[
 {
  "id": "2604.18176",
  "date": "2026-04-20",
  "title": "QuantumQA: Enhancing Scientific Reasoning via Physics-Consistent Dataset and Verification-Aware Reinforcement Learning",
  "scope_group": "other",
  "use": "reasoning-proof",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Qwen3-8B",
    "version": "Qwen3-8B",
    "role": "system"
   },
   {
    "name": "Meta-Llama-3.1-8B-Instruct",
    "version": "Meta-Llama-3.1-8B-Instruct",
    "role": "system"
   },
   {
    "name": "DeepSeek-R1-0528-Qwen3-8B",
    "version": "DeepSeek-R1-0528-Qwen3-8B",
    "role": "system"
   },
   {
    "name": "Qwen3-4B",
    "version": "Qwen3-4B",
    "role": "other"
   },
   {
    "name": "Qwen3-8B",
    "version": "Qwen3-8B",
    "role": "baseline"
   },
   {
    "name": "Meta-Llama-3.1-8B-Instruct",
    "version": "Meta-Llama-3.1-8B-Instruct",
    "role": "baseline"
   },
   {
    "name": "DeepSeek-R1-0528-Qwen3-8B",
    "version": "DeepSeek-R1-0528-Qwen3-8B",
    "role": "baseline"
   },
   {
    "name": "ChatGPT-5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "DeepSeek-R1-0528",
    "version": "DeepSeek-R1-0528",
    "role": "baseline"
   },
   {
    "name": "GLM-4.6",
    "version": "GLM-4.6",
    "role": "baseline"
   },
   {
    "name": "Qwen3-Max",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Kimi-K2-Thinking",
    "version": "Kimi-K2-Thinking",
    "role": "baseline"
   },
   {
    "name": "Qwen3-235B-A22B-Instruct",
    "version": "Qwen3-235B-A22B-Instruct",
    "role": "baseline"
   },
   {
    "name": "Skywork-Reward",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "DeepSeek-V3",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "DeepSeek-V3",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Qwen3-Max",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "ChatGPT-5",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "GPT-4o",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "GPT-4o",
    "version": null,
    "role": "other"
   },
   {
    "name": "Qwen2.5-1.5B",
    "version": "Qwen2.5-1.5B",
    "role": "data-generation"
   },
   {
    "name": "Qwen2.5-7B",
    "version": "Qwen2.5-7B",
    "role": "data-generation"
   },
   {
    "name": "Meta-Llama-3.1-8B-Instruct",
    "version": "Meta-Llama-3.1-8B-Instruct",
    "role": "data-generation"
   },
   {
    "name": "Qwen2.5-72B",
    "version": "Qwen2.5-72B",
    "role": "data-generation"
   },
   {
    "name": "Grok-3",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "ChatGPT-o1",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "Anthropic Claude 3.5 Sonnet",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Google Gemini 2.5 Pro",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Qwen3-Max",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Kimi-K2-Thinking",
    "version": "Kimi-K2-Thinking",
    "role": "judge"
   },
   {
    "name": "ChatGPT-5",
    "version": null,
    "role": "judge"
   },
   {
    "name": "ChatGPT",
    "version": null,
    "role": "writing"
   },
   {
    "name": "all-MiniLM-L6-v2",
    "version": "all-MiniLM-L6-v2",
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "glm-4.6",
   "newest_release": "2025-09-30",
   "newest_class": "open-weights",
   "lag_days": 202,
   "frontier_at_paper": "Claude Opus 4.7 (2026-04-16)",
   "flagship_releases_since": 13
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 73
  }
 },
 {
  "id": "2604.18805",
  "date": "2026-04-20",
  "title": "AI scientists produce results without reasoning scientifically",
  "scope_group": "other",
  "use": "agent-evaluation",
  "kind": "benchmark",
  "harness": "custom: Corral",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/lamalab-org/corral",
   "https://doi.org/10.5281/zenodo.19659851"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT-4o",
    "version": "gpt-4o-2024-08-06",
    "role": "system"
   },
   {
    "name": "Claude Sonnet 4.5",
    "version": "claude-sonnet-4-5-20250929",
    "role": "system"
   },
   {
    "name": "GPT-OSS",
    "version": "gpt-oss-120b",
    "role": "system"
   },
   {
    "name": "Claude Sonnet 4.5",
    "version": null,
    "role": "judge"
   },
   {
    "name": "ibm-research/MoLFormer-XL-both-10pct",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-4.5",
   "newest_release": "2025-09-29",
   "newest_class": "frontier-api",
   "lag_days": 203,
   "frontier_at_paper": "Claude Opus 4.7 (2026-04-16)",
   "flagship_releases_since": 13
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 189
  }
 },
 {
  "id": "2604.20304",
  "date": "2026-04-22",
  "title": "LLM-guided phase diagram construction through high-throughput experimentation",
  "scope_group": "materials-science",
  "use": "experimental-control",
  "kind": "application",
  "harness": "off-the-shelf: Claude Code; custom NIMO LLMEP",
  "harness_class": "off-the-shelf",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/NIMS-DA/nimo",
   "https://github.com/NIMS-DA/aLLoyM_pred",
   "https://huggingface.co/Playingyoyo/aLLoyM"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Opus 4.6",
    "version": "claude-opus-4-6",
    "role": "system"
   },
   {
    "name": "aLLoyM",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.6",
   "newest_release": "2026-02-05",
   "newest_class": "frontier-api",
   "lag_days": 76,
   "frontier_at_paper": "Claude Opus 4.7 (2026-04-16)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "alloys",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "alloys",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2604.20899",
  "date": "2026-04-21",
  "title": "Predicting Scale-Up of Metal-Organic Framework Syntheses with Large Language Models",
  "scope_group": "materials-science",
  "use": "property-prediction",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/zzhenglab/MOF-Scaleup/"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "GPT-4.1",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-4.1",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5.1",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "LLaMA-3-8B",
    "version": "Meta-Llama-3-8B-Instruct",
    "role": "baseline"
   },
   {
    "name": "DeepSeek-V3.2",
    "version": "deepseek-chat",
    "role": "baseline"
   },
   {
    "name": "BERT",
    "version": "bert-base-uncased",
    "role": "baseline"
   },
   {
    "name": "text-embedding-3-small",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4",
   "newest_release": "2026-03-05",
   "newest_class": "frontier-api",
   "lag_days": 47,
   "frontier_at_paper": "Claude Opus 4.7 (2026-04-16)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 45
  },
  "subject_codes": {
   "sol": {
    "subject": "porous-frameworks",
    "question": "method-capability"
   },
   "opus": {
    "subject": "porous-frameworks",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2604.21850",
  "date": "2026-04-23",
  "title": "OptiMat Alloys: a FAIR, living database of multi-principal element alloys enabled by a conversational agent",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "application",
  "harness": "off-the-shelf: AutoGen",
  "harness_class": "off-the-shelf",
  "base_models": 5,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "LLM dates not stated; database development/testing October 2025–April 2026",
  "code_urls": [
   "https://github.com/OptiMat-Chat/OptiMat-Alloys",
   "https://github.com/vladturlo/OptiMat-Chat"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GLM-4.5-Air",
    "version": null,
    "role": "system"
   },
   {
    "name": "MiMo-V2-Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-OSS-120B",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-4.1",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-OSS-20B",
    "version": null,
    "role": "system"
   },
   {
    "name": "gpt-oss:20b (Ollama)",
    "version": "gpt-oss:20b; MXFP4",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "mimo-v2-flash",
   "newest_release": "2025-12-16",
   "newest_class": "open-weights",
   "lag_days": 128,
   "frontier_at_paper": "GPT-5.5 Pro (2026-04-23)",
   "flagship_releases_since": 7
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 81
  },
  "subject_codes": {
   "sol": {
    "subject": "alloys",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "alloys",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2604.22571",
  "date": "2026-04-24",
  "title": "LARA: Validation-Driven Agentic Supercomputer Workflows for Atomistic Modeling",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GEMINI 2.5 Flash",
    "version": "2.5 Flash",
    "role": "system"
   },
   {
    "name": "Sonnet 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "gemini-embedding-001",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-4.6",
   "newest_release": "2026-02-17",
   "newest_class": "frontier-api",
   "lag_days": 66,
   "frontier_at_paper": "GPT-5.5 Pro (2026-04-23)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 42
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2604.22938",
  "date": "2026-04-24",
  "title": "Large language model-enabled automated data extraction for concrete materials informatics",
  "scope_group": "materials-science",
  "use": "literature-knowledge",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 17,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated; December 2025 dates pricing, not evaluation",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT-5",
    "version": "gpt-5-2025-08-07",
    "role": "system"
   },
   {
    "name": "GPT-4o",
    "version": "gpt-4o-2024-08-06",
    "role": "system"
   },
   {
    "name": "Claude Sonnet 4.5",
    "version": "anthropic/claude-sonnet-4.5",
    "role": "system"
   },
   {
    "name": "Claude Haiku 4.5",
    "version": "anthropic/claude-haiku-4.5",
    "role": "system"
   },
   {
    "name": "Gemini 2.5 Pro",
    "version": "google/gemini-2.5-pro",
    "role": "system"
   },
   {
    "name": "Gemini 2.5 Flash",
    "version": "google/gemini-2.5-flash",
    "role": "system"
   },
   {
    "name": "Grok 4",
    "version": "x-ai/grok-4",
    "role": "system"
   },
   {
    "name": "Grok 4 Fast",
    "version": "x-ai/grok-4-fast",
    "role": "system"
   },
   {
    "name": "Qwen3 Max",
    "version": "qwen/qwen3-max",
    "role": "system"
   },
   {
    "name": "Mistral Medium 3.1",
    "version": "mistralai/mistral-medium-3.1",
    "role": "system"
   },
   {
    "name": "gpt-oss-120b",
    "version": "openai/gpt-oss-120b",
    "role": "system"
   },
   {
    "name": "Gemma 3 27B",
    "version": "google/gemma-3-27b-it",
    "role": "system"
   },
   {
    "name": "Llama 4 Maverick",
    "version": "meta-llama/llama-4-maverick",
    "role": "system"
   },
   {
    "name": "Llama 3.3 70B",
    "version": "meta-llama/llama-3.3-70b-instruct",
    "role": "system"
   },
   {
    "name": "DeepSeek R1",
    "version": "deepseek/deepseek-r1-0528",
    "role": "system"
   },
   {
    "name": "DeepSeek V3.1",
    "version": "deepseek/deepseek-chat-v3.1",
    "role": "system"
   },
   {
    "name": "Qwen3 235B A22B",
    "version": "qwen/qwen3-235b-a22b-2507",
    "role": "system"
   },
   {
    "name": "ChatGPT",
    "version": null,
    "role": "writing"
   }
  ],
  "model_lag": {
   "newest_model": "claude-haiku-4.5",
   "newest_release": "2025-10-15",
   "newest_class": "frontier-api",
   "lag_days": 191,
   "frontier_at_paper": "GPT-5.5 Pro (2026-04-23)",
   "flagship_releases_since": 14
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 184
  },
  "subject_codes": {
   "sol": {
    "subject": "structural-composites",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "structural-composites",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2604.23758",
  "date": "2026-04-26",
  "title": "Agentic Fusion of Large Atomic and Language Models to Accelerate Superconductor Discovery",
  "scope_group": "materials-science",
  "use": "research-orchestration",
  "kind": "application",
  "harness": "off-the-shelf: OpenClaw",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/openclaw/openclaw",
   "https://github.com/atomistic-machine-learning/SchNet",
   "https://github.com/gasteigerjo/dimenet",
   "https://github.com/txie-93/cgcnn",
   "https://github.com/materialsvirtuallab/megnet",
   "https://github.com/usnistgov/alignn",
   "https://github.com/YKQ98/Matformer",
   "https://github.com/divelab/DIG"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": "5.4",
    "role": "data-generation"
   },
   {
    "name": "Opus-4.6",
    "version": "4.6",
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4",
   "newest_release": "2026-03-05",
   "newest_class": "frontier-api",
   "lag_days": 52,
   "frontier_at_paper": "GPT-5.5 Pro (2026-04-23)",
   "flagship_releases_since": 2
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 200
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-materials",
    "question": "targeted-design"
   },
   "opus": {
    "subject": "quantum-materials",
    "question": "targeted-design"
   }
  }
 },
 {
  "id": "2604.23821",
  "date": "2026-04-26",
  "title": "Accelerating Quantum Materials Characterization: Hybrid Active Learning for Autonomous Spin Wave Spectroscopy",
  "scope_group": "condensed-matter",
  "use": "experimental-control",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/usnistgov/tasai",
   "https://github.com/usnistgov/paper-tasai"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Opus 4.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini 3",
    "version": null,
    "role": "system"
   },
   {
    "name": "gpt-5.2-codex",
    "version": "gpt-5.2-codex",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.2-codex",
   "newest_release": "2025-12-18",
   "newest_class": "frontier-api",
   "lag_days": 129,
   "frontier_at_paper": "GPT-5.5 Pro (2026-04-23)",
   "flagship_releases_since": 7
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-materials",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-materials",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2604.24283",
  "date": "2026-04-27",
  "title": "AutoQResearch: LLM-Guided Closed-Loop Policy Search for Adaptive Variational Quantum Optimization",
  "scope_group": "adjacent-quantum",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/SMU-Quantum/autoqresearch"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.3-Codex",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.3-codex",
   "newest_release": "2026-02-05",
   "newest_class": "frontier-api",
   "lag_days": 81,
   "frontier_at_paper": "GPT-5.5 Pro (2026-04-23)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 20
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2604.25199",
  "date": "2026-04-28",
  "title": "Kohn-Sham Hamiltonian from Effective Field Theory: Quasiparticle Band Narrowing from Frozen Core Dynamics",
  "scope_group": "condensed-matter",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 31
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-materials",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "general",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2604.25884",
  "date": "2026-04-28",
  "title": "QCalEval: Benchmarking Vision-Language Models for Quantum Calibration Plot Understanding",
  "scope_group": "adjacent-quantum",
  "use": "data-interpretation",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "April 6, 2026: six closed-model API versions; other dates not stated",
  "code_urls": [
   "https://github.com/nvidia/QCalEval"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.4",
    "version": "gpt-5.4",
    "role": "baseline"
   },
   {
    "name": "GPT-5.4",
    "version": "gpt-5.4",
    "role": "judge"
   },
   {
    "name": "GPT-5.4",
    "version": "gpt-5.4",
    "role": "data-generation"
   },
   {
    "name": "GPT-5.4-Mini",
    "version": "gpt-5.4-mini",
    "role": "baseline"
   },
   {
    "name": "Gemini-3.1-Pro",
    "version": "gemini-3.1-pro-preview",
    "role": "baseline"
   },
   {
    "name": "Gemini-3.1-Pro",
    "version": "gemini-3.1-pro-preview",
    "role": "judge"
   },
   {
    "name": "Gemini-3.1-Pro",
    "version": "gemini-3.1-pro-preview",
    "role": "data-generation"
   },
   {
    "name": "Gemini-3.1-Flash-Lite",
    "version": "gemini-3.1-flash-lite-preview",
    "role": "baseline"
   },
   {
    "name": "Claude Opus 4.6",
    "version": "claude-opus-4-6",
    "role": "baseline"
   },
   {
    "name": "Claude Sonnet 4.6",
    "version": "claude-sonnet-4-6",
    "role": "baseline"
   },
   {
    "name": "Claude Haiku 4.5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.5-397B-A17B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.5-397B-A17B",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "Qwen3.5-122B-A10B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.5-35B-A3B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.5-27B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.5-9B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.5-9B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemma-4-31B-IT",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "InternVL3-78B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "InternVL3-38B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "MiniCPM-o-4.5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Kimi-VL-A3B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "NVIDIA Ising Calibration 1",
    "version": "Ising-Calibration-1-35B-A3B",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gemma-4-31b",
   "newest_release": "2026-04-02",
   "newest_class": "open-weights",
   "lag_days": 26,
   "frontier_at_paper": "GPT-5.5 Pro (2026-04-23)",
   "flagship_releases_since": 2
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 8,
   "bibliography_size": 55
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2604.27709",
  "date": "2026-04-30",
  "title": "Conditional Generative Models Enable Targeted Exploration of MAX Phase Design Space",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "application",
  "harness": "none",
  "harness_class": "none",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/j-swaine/Conditional-Generation-of-MAX-Phases"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "CrystaLLM-π",
    "version": null,
    "role": "system"
   },
   {
    "name": "baseline model without a condition vector",
    "version": null,
    "role": "baseline"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 47
  },
  "subject_codes": {
   "sol": {
    "subject": "functional-solids",
    "question": "method-capability"
   },
   "opus": {
    "subject": "functional-solids",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.00803",
  "date": "2026-05-01",
  "title": "Can Coding Agents Reproduce Findings in Computational Materials Science?",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "benchmark",
  "harness": "custom; off-the-shelf: Claude Code 2.1.78, Codex CLI 0.115.0; Claude Agent SDK 0.1.48",
  "harness_class": "off-the-shelf",
  "base_models": 4,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/JHU-CLSP/AutoMat"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Sonnet 4.6",
    "version": "claude-sonnet-4-6",
    "role": "system"
   },
   {
    "name": "Claude Opus 4.6",
    "version": "claude-opus-4-6",
    "role": "system"
   },
   {
    "name": "GPT-5.4 (reasoning: high)",
    "version": "gpt-5.4-2026-03-05",
    "role": "system"
   },
   {
    "name": "Kimi K2.5",
    "version": "kimi-k2.5",
    "role": "system"
   },
   {
    "name": "Claude Opus 4.6",
    "version": "claude-opus-4-6",
    "role": "baseline"
   },
   {
    "name": "LLM-based evaluator",
    "version": null,
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4",
   "newest_release": "2026-03-05",
   "newest_class": "frontier-api",
   "lag_days": 57,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 3,
   "bibliography_size": 45
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2605.03205",
  "date": "2026-05-04",
  "title": "From Knowledge to Action: Outcomes of the 2025 Large Language Model (LLM) Hackathon for Applications in Materials Science and Chemistry",
  "scope_group": "other",
  "use": "research-orchestration",
  "kind": "review-perspective",
  "harness": "custom; off-the-shelf: LangGraph, LangChain, CrewAI, Pydantic AI, Google ADK, CodeAct, NotebookLM; MCP integrations",
  "harness_class": "off-the-shelf",
  "base_models": 40,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "Not stated; hackathon September 11–12, 2025; later development reported (§67).",
  "code_urls": [
   "https://github.com/hansgundlach/ChemCodeBench",
   "https://github.com/MauricioCafiero/UMADock/tree/main",
   "https://github.com/Arthurns16/hacktoon2025",
   "https://github.com/kinzani/Inzani-Group/tree/main/LLM_Hackathon_2025",
   "https://github.com/dreamboat26/hackathon",
   "https://github.com/yicao-elina/LLM4Chem-Explainable-synthesis.git",
   "https://github.com/BigDFT-group/llm-hackathon-2025",
   "https://github.com/jdsanc/MixSense.git",
   "https://github.com/tonylifepix/AgentLearn",
   "https://github.com/aritraroy24/the-fullerene-factory",
   "https://github.com/BaosenZ/mach-interp-LLM-hackathon-2025.git",
   "https://github.com/aishahasim/ChemBot---AI-Assisted-Chemistry-Assistant",
   "https://github.com/deepaksaipendyala/OntoKG",
   "https://github.com/ncsu-llm-hackathon-materials-2025/MINT-LLM.git",
   "https://github.com/materialdigital/ontomapper",
   "https://github.com/dpalmer-anl/AtomBridge.git",
   "https://github.com/napoles-uach/GAINS",
   "https://github.com/argonne-lcf/ChemGraph",
   "https://github.com/epatyukova/llm2025-hackathon-CLUE",
   "https://github.com/ViktoriiaBaib/NEDD",
   "https://github.com/anky3733/Traj",
   "https://github.com/Awwal41/MaterialSim.git",
   "https://github.com/mo-alkubaish/SCARA",
   "https://github.com/msehabibur/gcxgc_peakcards",
   "https://github.com/datalab-org/guillemot",
   "https://github.com/Oscuro-Phoenix/xascribe",
   "https://github.com/mattiasutancykeln/Baker",
   "https://github.com/554181320angela-lang/Poly-predictor",
   "https://gitlab.com/heingroup/llm-hackathon-2025",
   "https://github.com/ivoryzh/MCP4SDL",
   "https://github.com/Naveen-R-M/F.A.D.E",
   "https://github.com/gopal-iyer/MatFOMGen",
   "https://github.com/chiku-parida/DFTPilot",
   "https://github.com/ndaelman-hu/parse-patrol",
   "https://github.com/ram123-debug/mat-chem-llm-hackathon.git",
   "https://github.com/schandy2211/scale",
   "https://github.com/sugannathan/L.A.R.A-LLMs-as-Aerogel-Research-Assistants/tree/main",
   "https://github.com/souvikta/ODEForge",
   "https://github.com/killiansheriff/MP-LLM",
   "https://github.com/ahaibel/mp-property-analogies",
   "https://github.com/ashikiut/CaMEL-RAG",
   "https://github.com/fpriante/SuperconLLM-Multi-Agent-Framework-for-Automating-Superconductivity-Knowledge-Extraction",
   "https://github.com/srustisain/mit-catalyze",
   "https://github.com/gourav-k/LLM4Causal.git",
   "https://github.com/Ahri111/ZEROMAT",
   "https://github.com/iarretche/Unstructured-Lab-Notebook-Assistant-ULNA-",
   "https://github.com/zakidotai/MuMMIE",
   "https://github.com/jaydenlee97/2025-LLM-Hackathon-for-Applications-in-Materials-and-Chemistry/tree/main/code",
   "https://github.com/SubramanyamSahoo/LLM-Hackathon-for-Applications-in-Materials-and-Chemistry-2025",
   "https://github.com/rafaelespinosacastaneda/Hackaton-for-Applications-in-Materials-Science-and-Chemistry",
   "https://github.com/CGruich/RedoxFlow",
   "https://github.com/guannant/LLM_auto_featurization",
   "https://github.com/truptimohanty/MAGE",
   "https://github.com/abir0/dft-agent",
   "https://github.com/JSR-ISM-Smart-Chemistry-Lab/Titanarium",
   "https://github.com/icn2-ai/m2snano-llm",
   "https://github.com/schwallergroup/MDAgent",
   "https://github.com/MatAgentHub/crystalk",
   "https://github.com/MUzair20/spectrobot",
   "https://github.com/estefaniavazquez/MindMesh",
   "https://github.com/alexchen5/syntheseek",
   "https://github.com/ruiding-uchicago/auto_CPT_uma_simul",
   "https://github.com/FAIRmat-NFDI/nomad-bot-rag-docs-discord",
   "https://github.com/pagel-s/MIDAS.git",
   "https://github.com/schwallergroup/llm_adsorbate",
   "https://github.com/ndharms/peptide-agent",
   "https://github.com/AzizahAlq/MaterialMind",
   "https://huggingface.co/Abbasaabdul/AI_ChemTutor/tree/main",
   "https://github.com/MJRaei/CrystaLenz",
   "https://github.com/HassanHarb92/ACME",
   "https://github.com/staradutt/HEAquery/tree/main",
   "https://github.com/devanshu-777/PackSynth",
   "https://github.com/sptiwari/ExpAlign_paper",
   "https://github.com/sbanik2/QSPHAgents",
   "https://github.com/RishikeshMagar/GPT-OSS-MAT",
   "https://github.com/abdulazizashy5/Alhaytham_Eye",
   "https://github.com/puppalasaisrikar/VERA",
   "https://github.com/Areeb2735/MaterEase-Your-AI-Powered-Materials-Science-Assistant",
   "https://github.com/AI4ChemS/ChemUnityQA",
   "https://github.com/HallucinatingStrikeTeam/AESOP",
   "https://github.com/darianSmalley/explain-that-automation",
   "https://github.com/DIYANAPV/llm-hackathon-MS-and-Chem-2025/tree/main",
   "https://github.com/AntoRoyan/MOF-Genie",
   "https://github.com/hspark1212/sky",
   "https://huggingface.co/cafierom",
   "https://huggingface.co/datasets/aritraroy24/optimized_conformers_dataset",
   "https://huggingface.co/umairbinmansoor/finetuned_model/tree/main",
   "https://github.com/olivettigroup/MatKG",
   "https://github.com/argonne-lcf/ChemGraph/releases/tag/v2025.09.12",
   "https://github.com/ahaibel/mp-property-analogies/blob/main/hackathon_submission_resources/mae_across_methods_properties.csv"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Llama 3.1-8B",
    "version": null,
    "role": "other"
   },
   {
    "name": "Llama 3.3-70B",
    "version": null,
    "role": "other"
   },
   {
    "name": "GPT-4o mini",
    "version": null,
    "role": "other"
   },
   {
    "name": "Qwen-3-32B",
    "version": null,
    "role": "other"
   },
   {
    "name": "gpt-oss-120b",
    "version": null,
    "role": "other"
   },
   {
    "name": "Phi-4-mini-instruct",
    "version": "Phi-4-mini-instruct",
    "role": "system"
   },
   {
    "name": "mistral-large:123b-instruct-2407-q4_0",
    "version": "mistral-large:123b-instruct-2407-q4_0",
    "role": "system"
   },
   {
    "name": "gpt-oss:120b",
    "version": "gpt-oss:120b",
    "role": "system"
   },
   {
    "name": "Gemini",
    "version": null,
    "role": "system"
   },
   {
    "name": "ChatGPT",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Gemini",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Qwen2.5-7B",
    "version": "Qwen2.5-7B",
    "role": "system"
   },
   {
    "name": "DeepSeek model (V3:together)",
    "version": "V3:together",
    "role": "system"
   },
   {
    "name": "Gemini-2.5-Flash",
    "version": "Gemini-2.5-Flash",
    "role": "system"
   },
   {
    "name": "Qwen-3-235B",
    "version": null,
    "role": "system"
   },
   {
    "name": "DeepSeek-V3.2-Exp",
    "version": "DeepSeek-V3.2-Exp",
    "role": "system"
   },
   {
    "name": "Ether0",
    "version": null,
    "role": "other"
   },
   {
    "name": "DeepSeek-R1",
    "version": null,
    "role": "other"
   },
   {
    "name": "Gemini-2.5-Pro",
    "version": null,
    "role": "other"
   },
   {
    "name": "Mistral-Small-24B-Instruct-2501",
    "version": "Mistral-Small-24B-Instruct-2501",
    "role": "other"
   },
   {
    "name": "DeepSeek-R1",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "Gemini-2.5-Flash",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "GPT-5",
    "version": null,
    "role": "other"
   },
   {
    "name": "BioGPT",
    "version": "microsoft/biogpt",
    "role": "system"
   },
   {
    "name": "sentence-transformers/all-MiniLM-L6-v2",
    "version": "sentence-transformers/all-MiniLM-L6-v2",
    "role": "other"
   },
   {
    "name": "gpt-oss 20B",
    "version": "gpt-oss-20b",
    "role": "system"
   },
   {
    "name": "Llama3",
    "version": null,
    "role": "system"
   },
   {
    "name": "Mistral",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5",
    "version": "GPT-5",
    "role": "system"
   },
   {
    "name": "Gemini-2.5-Pro",
    "version": "Gemini-2.5-Pro",
    "role": "system"
   },
   {
    "name": "gpt-5-mini-2025-08-07",
    "version": "gpt-5-mini-2025-08-07",
    "role": "system"
   },
   {
    "name": "ChatGPT-4-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude",
    "version": null,
    "role": "other"
   },
   {
    "name": "T5-Transformer",
    "version": null,
    "role": "system"
   },
   {
    "name": "ChatGPT 5 pro",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "BGE-large-en",
    "version": null,
    "role": "other"
   },
   {
    "name": "text-embedding-3-large",
    "version": "text-embedding-3-large",
    "role": "other"
   },
   {
    "name": "GPT-4",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "MatSciBERT",
    "version": null,
    "role": "other"
   },
   {
    "name": "LLaMat",
    "version": null,
    "role": "system"
   },
   {
    "name": "ChatGPT (GPT-5.1)",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini-2.0-Flash",
    "version": "Gemini-2.0-Flash",
    "role": "system"
   },
   {
    "name": "gemini-2.5-flash-lite",
    "version": "gemini-2.5-flash-lite",
    "role": "system"
   },
   {
    "name": "gemini-1.5-flash",
    "version": "gemini-1.5-flash",
    "role": "system"
   },
   {
    "name": "GPT-5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5-mini",
    "version": "GPT-5-mini",
    "role": "system"
   },
   {
    "name": "gpt-4o-mini",
    "version": "gpt-4o-mini",
    "role": "system"
   },
   {
    "name": "gpt-4.1-mini",
    "version": "gpt-4.1-mini",
    "role": "system"
   },
   {
    "name": "LLaMA 3.2-3B",
    "version": "LLaMA-3.2-3B",
    "role": "system"
   },
   {
    "name": "Qwen-3-235b-a22b-instruct-2507",
    "version": "Qwen-3-235b-a22b-instruct-2507",
    "role": "system"
   },
   {
    "name": "Claude Sonnet 4",
    "version": "claude-sonnet-4-20250514",
    "role": "system"
   },
   {
    "name": "GPT-4o",
    "version": "GPT-4o",
    "role": "system"
   },
   {
    "name": "Gemini 2.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-4",
    "version": "GPT-4",
    "role": "system"
   },
   {
    "name": "LLMProp",
    "version": null,
    "role": "other"
   },
   {
    "name": "Whisper",
    "version": null,
    "role": "other"
   },
   {
    "name": "llama-4-maverick-17b-128e-instruct",
    "version": "llama-4-maverick-17b-128e-instruct",
    "role": "system"
   },
   {
    "name": "qwen3-235b-a22b-thinking-2507",
    "version": "qwen3-235b-a22b-thinking-2507",
    "role": "system"
   },
   {
    "name": "Qwen-3 32B",
    "version": "Qwen-3-32B",
    "role": "system"
   },
   {
    "name": "ChatGPT",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "Perplexity",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "LLaMA 3.1 (8B)",
    "version": "LLaMA-3.1-8B",
    "role": "system"
   },
   {
    "name": "GP-MoLFormer-Uniq",
    "version": "GP-MoLFormer-Uniq",
    "role": "system"
   },
   {
    "name": "Mistral-7B-v0.3",
    "version": "Mistral-7B-v0.3",
    "role": "system"
   },
   {
    "name": "GPT-4.1",
    "version": "GPT-4.1",
    "role": "system"
   },
   {
    "name": "GPT-4.1",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "Haiku 3.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Opus 4.1",
    "version": null,
    "role": "system"
   },
   {
    "name": "Sonnet 4.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Nomic-Embed-Text",
    "version": null,
    "role": "other"
   },
   {
    "name": "T5-base",
    "version": "T5-base",
    "role": "system"
   },
   {
    "name": "ChemT5",
    "version": null,
    "role": "system"
   },
   {
    "name": "TinyLLaMA",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-3.5",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "Socratic AI reasoning model",
    "version": null,
    "role": "system"
   },
   {
    "name": "BART",
    "version": null,
    "role": "other"
   },
   {
    "name": "GPT-2",
    "version": "GPT-2",
    "role": "system"
   },
   {
    "name": "LLaMA-2 7B",
    "version": "LLaMA-2-7B",
    "role": "baseline"
   },
   {
    "name": "LLaMA-2 13B",
    "version": "LLaMA-2-13B",
    "role": "baseline"
   },
   {
    "name": "LLaMA-2 70B",
    "version": "LLaMA-2-70B",
    "role": "baseline"
   },
   {
    "name": "Claude-class configuration",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5-mini",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Claude-Sonnet-4",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini-2.5-Pro",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Grok-4",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "LLM-as-judge",
    "version": null,
    "role": "judge"
   },
   {
    "name": "code llama",
    "version": null,
    "role": "system"
   },
   {
    "name": "O3",
    "version": null,
    "role": "system"
   },
   {
    "name": "O4-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "Base LLM",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "OpenCLIP",
    "version": null,
    "role": "other"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 4,
   "bibliography_size": 340
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "ai-workflow-perspective"
   },
   "opus": {
    "subject": "general",
    "question": "ai-workflow-perspective"
   }
  }
 },
 {
  "id": "2605.03515",
  "date": "2026-05-05",
  "title": "Scale-Dependent Input Representation and Confidence Estimation for LLMs in Materials Property Prediction",
  "scope_group": "materials-science",
  "use": "property-prediction",
  "kind": "application",
  "harness": "none",
  "harness_class": "none",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/shu-ozawa/crystal-llm-representations",
   "https://huggingface.co/ozashu/crystal-llm-representations"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Llama-3.2-1B-Instruct",
    "version": "Llama-3.2-1B-Instruct",
    "role": "system"
   },
   {
    "name": "Llama-3.1-8B-Instruct",
    "version": "Llama-3.1-8B-Instruct",
    "role": "system"
   },
   {
    "name": "LLM-Prop",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "MatBERT",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "llama-3.2-1b",
   "newest_release": "2024-09-25",
   "newest_class": "open-weights",
   "lag_days": 587,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 31
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 40
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2605.06839",
  "date": "2026-05-07",
  "title": "LLM-Guided Open Hypothesis Learning from Autonomous Scanning Probe Microscopy Experiments",
  "scope_group": "condensed-matter",
  "use": "data-interpretation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Slautin/2026_domain_growth_hypothesis_learning",
   "https://github.com/RichardLiuCoding/aespm"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Gemini",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-based deep-research pipeline",
    "version": null,
    "role": "data-generation"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "functional-solids",
    "question": "method-capability"
   },
   "opus": {
    "subject": "functional-solids",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.07145",
  "date": "2026-05-08",
  "title": "Fine-tuning a vision-language model for fracture-surface morphology recognition",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "application",
  "harness": "none",
  "harness_class": "none",
  "base_models": 3,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://doi.org/10.6084/m9.figshare.32086971"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen3-VL-32B-Instruct",
    "version": "Qwen/Qwen3-VL-32B-Instruct",
    "role": "system"
   },
   {
    "name": "GPT-5.2-Reasoning",
    "version": "gpt-5.2-2025-12-11",
    "role": "data-generation"
   },
   {
    "name": "GPT-5.4-Reasoning",
    "version": "gpt-5.4-2026-03-05",
    "role": "judge"
   },
   {
    "name": "GPT-5.5-Reasoning",
    "version": "gpt-5.5-2026-04-23",
    "role": "system"
   },
   {
    "name": "Gemini 3.1 Pro-Reasoning",
    "version": "gemini-3.1-pro-preview",
    "role": "system"
   },
   {
    "name": "Llama-3.2-11B-Vision",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GLM-4.6V",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "InternVL3.5",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 15,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.08255",
  "date": "2026-05-07",
  "title": "Can LLMs Predict Polymer Physics Just by Reading Synthesis and Processing Prose?",
  "scope_group": "materials-science",
  "use": "property-prediction",
  "kind": "method",
  "harness": "none",
  "harness_class": "none",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen3.5-9B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3.5-9B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Claude 3 Opus 4.6",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "qwen3.5-9b",
   "newest_release": "2026-03-02",
   "newest_class": "open-weights",
   "lag_days": 66,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 25
  },
  "subject_codes": {
   "sol": {
    "subject": "molecular-soft",
    "question": "method-capability"
   },
   "opus": {
    "subject": "molecular-soft",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.10120",
  "date": "2026-05-11",
  "title": "MicroWorld: Empowering Multimodal Large Language Models to Bridge the Microscopic Domain Gap with Multimodal Attribute Graph",
  "scope_group": "other",
  "use": "scientific-qa",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 6,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/ieellee/MicroWorld"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen3-VL-2B-Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-VL-4B-Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-VL-8B-Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "InternVL3.5-2B-Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "InternVL3.5-4B-Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "InternVL3.5-8B-Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "Qwen3-VL-Embedding",
    "version": null,
    "role": "other"
   },
   {
    "name": "o1",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Claude Sonnet 4.5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "o4-mini",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "o3",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "qwen3-vl-2b",
   "newest_release": "2025-10-21",
   "newest_class": "open-weights",
   "lag_days": 202,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 15
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 50
  }
 },
 {
  "id": "2605.10327",
  "date": "2026-05-11",
  "title": "SCALAR: A Neurosymbolic Framework for Automated Conjecture and Reasoning in Quantum Circuit Analysis",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/sfeeney1897/SCALAR"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Sonnet 4.6",
    "version": "4.6",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-4.6",
   "newest_release": "2026-02-17",
   "newest_class": "frontier-api",
   "lag_days": 83,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 32
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.13618",
  "date": "2026-05-13",
  "title": "OpenAaaS: An Open Agent-as-a-Service Framework for Distributed Materials-Informatics Research",
  "scope_group": "materials-science",
  "use": "research-orchestration",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Wolido/OpenAaaS"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Kimi-K2.6",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 20,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 51
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "general",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2605.13911",
  "date": "2026-05-13",
  "title": "Rongzai agent: A Large Language Model-Based Autonomous Assistant for Rietveld Refinement of Neutron Diffraction Data",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "application",
  "harness": "off-the-shelf: Open Dr.Sai; custom GSAS-II/FastMCP interface",
  "harness_class": "off-the-shelf",
  "base_models": 4,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/hepai-lab/drsai"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-4.1",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5",
    "version": null,
    "role": "system"
   },
   {
    "name": "MiniMax-m2.7",
    "version": null,
    "role": "system"
   },
   {
    "name": "Scienceone-s1-base-ultra",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "minimax-m2.7",
   "newest_release": "2026-03-18",
   "newest_class": "open-weights",
   "lag_days": 56,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 2,
   "bibliography_size": 19
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "general",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2605.14344",
  "date": "2026-05-14",
  "title": "CrystalReasoner: Reasoning and RL for Property-Conditioned Crystal Structure Generation",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/wyy603/CrystalReasoner/"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Qwen2.5-3B",
    "version": null,
    "role": "system"
   },
   {
    "name": "CrystalTextLLM",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "PLAID++ Wyckoff Base",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "qwen2.5-3b",
   "newest_release": "2024-09-19",
   "newest_class": "open-weights",
   "lag_days": 602,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 31
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 2,
   "bibliography_size": 49
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.14527",
  "date": "2026-05-14",
  "title": "Lang2MLIP: End-to-End Language-to-Machine Learning Interatomic Potential Development with Autonomous Agentic Workflows",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "off-the-shelf: Claude Agent SDK",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Opus 4.6",
    "version": "4.6",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.6",
   "newest_release": "2026-02-05",
   "newest_class": "frontier-api",
   "lag_days": 98,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 7
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 39
  },
  "subject_codes": {
   "sol": {
    "subject": "battery-materials",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "general",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2605.14671",
  "date": "2026-05-14",
  "title": "Agentic Design of Compositional Descriptors via Autoresearch for Materials Science Applications",
  "scope_group": "materials-science",
  "use": "property-prediction",
  "kind": "method",
  "harness": "custom: Automat; off-the-shelf: OpenAI Codex",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/m-cobelli/automat",
   "https://github.com/karpathy/autoresearch"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.5",
    "version": "GPT-5.5",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 21,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 49
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.15227",
  "date": "2026-05-13",
  "title": "NIMO Controller: a self-driving laboratory orchestrator based on the Model Context Protocol",
  "scope_group": "other",
  "use": "experimental-control",
  "kind": "method",
  "harness": "off-the-shelf: Pydantic AI",
  "harness_class": "off-the-shelf",
  "base_models": 8,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/NIMS-DA/nimo-controller"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "claude-opus-5",
    "version": null,
    "role": "system"
   },
   {
    "name": "gpt-5.6-luna",
    "version": null,
    "role": "system"
   },
   {
    "name": "gpt-oss:120b",
    "version": null,
    "role": "system"
   },
   {
    "name": "gpt-oss:20b",
    "version": null,
    "role": "system"
   },
   {
    "name": "qwen3.8:27b",
    "version": null,
    "role": "system"
   },
   {
    "name": "qwen3.6:35b",
    "version": null,
    "role": "system"
   },
   {
    "name": "gemma4:31b",
    "version": null,
    "role": "system"
   },
   {
    "name": "gemma4:e4b",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "qwen3.8-27b",
   "newest_release": "2026-08-14",
   "newest_class": "open-weights",
   "lag_days": -93,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 0
  },
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 0,
   "bibliography_size": 25
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "general",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2605.15625",
  "date": "2026-05-15",
  "title": "ColPackAgent: Agent-Skill-Guided Hard-Particle Monte Carlo Workflows for Colloidal Packing",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 10,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "Benchmark not stated; SI B report dated 2026-05-05.",
  "code_urls": [
   "https://github.com/ljding94/ColPackAgent"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen3-Next 80B Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Opus 4.7",
    "version": "4.7",
    "role": "system"
   },
   {
    "name": "Gemini 2.5 Flash",
    "version": "2.5",
    "role": "system"
   },
   {
    "name": "claude haiku-4.5",
    "version": "4.5",
    "role": "system"
   },
   {
    "name": "DeepSeek V3.2",
    "version": "V3.2",
    "role": "system"
   },
   {
    "name": "gemini-3.1 pro-preview",
    "version": "3.1 pro-preview",
    "role": "system"
   },
   {
    "name": "gpt-oss-120b",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Opus 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "Sonnet 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "gpt-5.4",
    "version": "5.4",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.7",
   "newest_release": "2026-04-16",
   "newest_class": "frontier-api",
   "lag_days": 29,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 2
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 36
  },
  "subject_codes": {
   "sol": {
    "subject": "molecular-soft",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "molecular-soft",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2605.15842",
  "date": "2026-05-15",
  "title": "Reconstructing temporal multi-relational firm networks at scale using large language models. The case of the semiconductor industry",
  "scope_group": "other",
  "use": "literature-knowledge",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/georgetown-cset/eto-chip-explorer"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-4o-mini",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-4o-mini",
   "newest_release": "2024-07-18",
   "newest_class": "frontier-api",
   "lag_days": 666,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 31
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 47
  }
 },
 {
  "id": "2605.16191",
  "date": "2026-05-15",
  "title": "Optimized Three-Dimensional Photovoltaic Structures with LLM guided Tree Search",
  "scope_group": "other",
  "use": "design-optimization",
  "kind": "application",
  "harness": "off-the-shelf: ERA and Google AntiGravity",
  "harness_class": "off-the-shelf",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/feranick/3dpv"
  ],
  "ai_use_statement": false,
  "models": [],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 3
  },
  "subject_codes": {
   "sol": {
    "subject": "semiconductor-optical",
    "question": "method-capability"
   },
   "opus": {
    "subject": "semiconductor-optical",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.17254",
  "date": "2026-05-17",
  "title": "CatalyticMLLM: A Graph-Text Multimodal Large Language Model for Catalytic Materials",
  "scope_group": "materials-science",
  "use": "property-prediction",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/atomicarchitects/equiformer",
   "https://github.com/atomicarchitects/equiformer_v2"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen2.5-VL / Qwen2.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Llama3.2",
    "version": "3.2",
    "role": "system"
   },
   {
    "name": "CatBERTa",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "RoBERTa",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GAP-CatBERTa",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "MatterChat",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "QE-Catalytic",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "CrysText",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "CrysText-RL",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "llama-3.2",
   "newest_release": "2024-09-25",
   "newest_class": "open-weights",
   "lag_days": 599,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 31
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 39
  },
  "subject_codes": {
   "sol": {
    "subject": "catalysts",
    "question": "method-capability"
   },
   "opus": {
    "subject": "catalysts",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.17675",
  "date": "2026-05-17",
  "title": "Bridging the Gap on AI-Assisted Scientific Software Development Through Transparency and Traceability",
  "scope_group": "other",
  "use": "code-algorithm-generation",
  "kind": "proposal",
  "harness": "off-the-shelf: Claude-code; Codex",
  "harness_class": "off-the-shelf",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/idaholab/TMAP8",
   "https://github.com/idaholab/TMAP8/pull/397",
   "https://github.com/idaholab/TMAP8/pull/402",
   "https://github.com/idaholab/TMAP8/pull/409"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Opus 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "ChatGPT 5.4",
    "version": "5.4",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4",
   "newest_release": "2026-03-05",
   "newest_class": "frontier-api",
   "lag_days": 73,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 61
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "ai-workflow-perspective"
   },
   "opus": {
    "subject": "general",
    "question": "ai-workflow-perspective"
   }
  }
 },
 {
  "id": "2605.18407",
  "date": "2026-05-18",
  "title": "Qumus: Realization of An Embodied AI Quantum Material Experimentalist",
  "scope_group": "condensed-matter",
  "use": "experimental-control",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 6,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.4",
    "version": "5.4",
    "role": "system"
   },
   {
    "name": "Gemini-3.1-pro",
    "version": "3.1-pro",
    "role": "system"
   },
   {
    "name": "Claude-Sonnet-4-6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "Grok-4.1-Fast-Reasoning",
    "version": "4.1-Fast-Reasoning",
    "role": "system"
   },
   {
    "name": "Qwen-Plus",
    "version": null,
    "role": "system"
   },
   {
    "name": "Deepseek-Chat",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4",
   "newest_release": "2026-03-05",
   "newest_class": "frontier-api",
   "lag_days": 74,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "semiconductor-optical",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "semiconductor-optical",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2605.20242",
  "date": "2026-05-18",
  "title": "LEAP: A closed-loop framework for perovskite precursor additive discovery",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/WD928/LEAP"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Perovskite-RL",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-32B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "gemini-3-flash",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "DeepSeek-V3.2",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "llama-3.3-70b-instruct",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "gemini-3-flash",
   "newest_release": "2025-12-17",
   "newest_class": "frontier-api",
   "lag_days": 152,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 8
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 58
  },
  "subject_codes": {
   "sol": {
    "subject": "semiconductor-optical",
    "question": "method-capability"
   },
   "opus": {
    "subject": "semiconductor-optical",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.20416",
  "date": "2026-05-19",
  "title": "Miller-Index-Based Latent Crystallographic Fracture Plane Reasoning and generation with Vision-Language Models",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 14
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2605.20495",
  "date": "2026-05-19",
  "title": "A Human-in-the-Loop Framework for Efficient Prompt Selection in Microscopy Vision-Language Models",
  "scope_group": "other",
  "use": "data-interpretation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://abhiram-kandiyana.github.io/APT/",
   "https://github.com/openai/tiktoken"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-4o",
    "version": null,
    "role": "system"
   },
   {
    "name": "Sentence-Transformer",
    "version": null,
    "role": "other"
   },
   {
    "name": "CLIP",
    "version": "ViT-B/16",
    "role": "other"
   },
   {
    "name": "BioMedCLIP",
    "version": "ViT-B/16; PubMedBERT text encoder",
    "role": "other"
   },
   {
    "name": "MedSigLIP",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-4o",
   "newest_release": "2024-05-13",
   "newest_class": "frontier-api",
   "lag_days": 736,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 31
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 34
  }
 },
 {
  "id": "2605.23934",
  "date": "2026-04-24",
  "title": "Practical Quantum CIM Empowerment via All-Domestic-Core Agentic Large Model",
  "scope_group": "adjacent-quantum",
  "use": "simulation-workflow",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/wrlengo/QUBO-VAE",
   "https://github.com/langchain-ai"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Doubao-1.5-Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen2.5-14B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini-3-Pro",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gemini-3-pro",
   "newest_release": "2025-11-18",
   "newest_class": "frontier-api",
   "lag_days": 157,
   "frontier_at_paper": "GPT-5.5 Pro (2026-04-23)",
   "flagship_releases_since": 11
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 52
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2605.24002",
  "date": "2026-05-18",
  "title": "Harnessing AtomisticSkills for Agentic Atomistic Research",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom: AtomisticSkills; off-the-shelf: Google Antigravity/OpenClaw demonstrations",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/learningmatter-mit/AtomisticSkills"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.4-mini",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "GPT-5.4-mini",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Claude Sonnet 4.6 (Thinking)",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "unnamed VLM",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4-mini",
   "newest_release": "2026-03-17",
   "newest_class": "frontier-api",
   "lag_days": 62,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 4,
   "bibliography_size": 125
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "general",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2605.26021",
  "date": "2026-05-25",
  "title": "Toward General Quantum Control with Physics-Informed Large Language Models",
  "scope_group": "adjacent-quantum",
  "use": "design-optimization",
  "kind": "method",
  "harness": "off-the-shelf: SDE-Harness; LiteLLM",
  "harness_class": "off-the-shelf",
  "base_models": 4,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/BerriAI/litellm"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Sonnet 4.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini 3.0 Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-OSS-120B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen 3.5 Plus",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "qwen3.5-plus",
   "newest_release": "2026-02-15",
   "newest_class": "frontier-api",
   "lag_days": 99,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 56
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.26179",
  "date": "2026-05-25",
  "title": "AutoDFT: A Closed-Loop Multi-Agent Framework for Autonomous DFT Calculations",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.2",
    "version": "GPT-5.2",
    "role": "system"
   },
   {
    "name": "Claude Sonnet 4.6",
    "version": "Claude Sonnet 4.6",
    "role": "system"
   },
   {
    "name": "Gemini 3.1 Pro Preview",
    "version": "Gemini 3.1 Pro Preview",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gemini-3.1-pro",
   "newest_release": "2026-02-19",
   "newest_class": "frontier-api",
   "lag_days": 95,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 3,
   "bibliography_size": 30
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "general",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2605.26741",
  "date": "2026-05-26",
  "title": "MatFormBench: A Benchmarking Evaluation Framework for Target-Driven Materials Formulation",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/DeepVerse/MatFormBench"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "DeepSeek-v4-flash",
    "version": "v4-flash",
    "role": "baseline"
   },
   {
    "name": "GLM-5.1",
    "version": "5.1",
    "role": "baseline"
   },
   {
    "name": "Kimi K2.6",
    "version": "K2.6",
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "deepseek-v4-flash",
   "newest_release": "2026-04-24",
   "newest_class": "open-weights",
   "lag_days": 32,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 56
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2605.27210",
  "date": "2026-05-26",
  "title": "Qiskit QuantumKatas: Adapting Microsoft's Quantum Computing exercises for LLM evaluation",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/qiskit-community/Qiskit-QuantumKatas/"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Opus 4.7",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Claude Sonnet 4.6",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Claude Haiku 4.5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5.3-Codex",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini 3.1 Pro",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Mistral Large 3",
    "version": "Mistral-Large-3-675B-Instruct-2512",
    "role": "baseline"
   },
   {
    "name": "Llama 4 Maverick",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-OSS-120B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Llama 4 Scout",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemma 4 31B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Granite 4.1 30B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemma 4 26B-A4B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Mistral Small 3.2 24B",
    "version": "Mistral-Small-3.2-24B-Instruct-2506",
    "role": "baseline"
   },
   {
    "name": "GPT-OSS-20B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Granite 4.1 8B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Claude Code",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "Qiskit Code Assistant",
    "version": null,
    "role": "data-generation"
   }
  ],
  "model_lag": {
   "newest_model": "granite-4.1-8b",
   "newest_release": "2026-04-29",
   "newest_class": "open-weights",
   "lag_days": 27,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 2,
   "bibliography_size": 54
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2605.28487",
  "date": "2026-05-27",
  "title": "ProvMind: Provenance-grounded reasoning for materials synthesis",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/ZHymLumine/MatProcBench"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen/Qwen2.5-7B-Instruct",
    "version": "Qwen2.5-7B-Instruct",
    "role": "system"
   },
   {
    "name": "Qwen/Qwen2.5-7B-Instruct",
    "version": "Qwen2.5-7B-Instruct",
    "role": "baseline"
   },
   {
    "name": "all-mpnet-base-v2",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "qwen2.5-7b",
   "newest_release": "2024-09-19",
   "newest_class": "open-weights",
   "lag_days": 615,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 31
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 39
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.29179",
  "date": "2026-05-27",
  "title": "Sustainable Metal-Organic Framework Water Harvesters in the Artificial Intelligence Era",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "review-perspective",
  "harness": "none",
  "harness_class": "none",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [],
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "porous-frameworks",
    "question": "literature-findings"
   },
   "opus": {
    "subject": "porous-frameworks",
    "question": "literature-findings"
   }
  }
 },
 {
  "id": "2605.29394",
  "date": "2026-05-28",
  "title": "EvoMD-LLM: Learning the Language of Species Evolution in Reactive Molecular Dynamics",
  "scope_group": "other",
  "use": "data-interpretation",
  "kind": "method",
  "harness": "none",
  "harness_class": "none",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/unslothai/unsloth",
   "https://github.com/open-compass/opencompass"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Llama 3.1 8B Instruct",
    "version": "3.1",
    "role": "system"
   },
   {
    "name": "Llama-3.1",
    "version": "3.1",
    "role": "baseline"
   },
   {
    "name": "ChemDFM",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3",
    "version": null,
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "llama-3.1-8b",
   "newest_release": "2024-07-23",
   "newest_class": "open-weights",
   "lag_days": 674,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 32
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 35
  }
 },
 {
  "id": "2605.29446",
  "date": "2026-05-28",
  "title": "CrystalXRD-Bench: Benchmarking Vision-Language Models for XRD Peak Indexing Across Diverse Crystalline Materials",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "benchmark",
  "harness": "none",
  "harness_class": "none",
  "base_models": 7,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://huggingface.co/datasets/xiaodu-ali/CrystalXRD-Bench"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini 3.1 Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini 3.0 Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3.6-Plus",
    "version": null,
    "role": "system"
   },
   {
    "name": "Doubao-Seed-2.0",
    "version": null,
    "role": "system"
   },
   {
    "name": "Kimi-K2.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Opus 4.6",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "qwen3.6-plus",
   "newest_release": "2026-04-02",
   "newest_class": "frontier-api",
   "lag_days": 56,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 2,
   "bibliography_size": 26
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2605.29555",
  "date": "2026-05-28",
  "title": "From Blind Guess to Informed Judgment: Teaching LLMs to Evaluate Materials by Building Knowledge-Augmented Preference Signals",
  "scope_group": "materials-science",
  "use": "property-prediction",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 6,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/yuyouyu32/MaterEval"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Hunyuan-Instruct (1.8B)",
    "version": null,
    "role": "system"
   },
   {
    "name": "Hunyuan-Instruct (4B)",
    "version": null,
    "role": "system"
   },
   {
    "name": "Hunyuan-7B-Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-Instruct (1.7B)",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-Instruct (4B)",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-8B-Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5",
    "version": null,
    "role": "judge"
   },
   {
    "name": "DeepSeek-R1",
    "version": null,
    "role": "judge"
   },
   {
    "name": "GPT-4o",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Claude-Sonnet-4.5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini-2.5-Pro",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-4.5",
   "newest_release": "2025-09-29",
   "newest_class": "frontier-api",
   "lag_days": 241,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 16
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 33
  },
  "subject_codes": {
   "sol": {
    "subject": "alloys",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2605.29833",
  "date": "2026-05-28",
  "title": "OmniMatBench: A Human-Calibrated Multimodal Reasoning Benchmark Across 19 Materials Science Subfields",
  "scope_group": "materials-science",
  "use": "scientific-qa",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Opus 4.7",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini 3.1 Pro",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.6 Plus",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Grok-4.3",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "o3",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "o4-mini",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-4o",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemma-4-31B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3-VL-235B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Intern-S1 Pro",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GLM-4.6V",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Llama-4 Maverick",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "LLaMA-3.2-11B-Vision-Instruct",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3-VL-8B-Instruct",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3-VL-8B-Thinking",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini 2.5 Flash",
    "version": null,
    "role": "judge"
   },
   {
    "name": "unnamed LLM verifier",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "grok-4.3",
   "newest_release": "2026-04-30",
   "newest_class": "frontier-api",
   "lag_days": 28,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 31
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2605.30358",
  "date": "2026-04-28",
  "title": "QASM-Eval: A Dataset to Train and Evaluate LLMs on OpenQASM-3 Beyond Quantum Circuits",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "benchmark",
  "harness": "off-the-shelf: OpenCode (GPT agent baseline); custom evaluation",
  "harness_class": "off-the-shelf",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/fuzhenxiao/QASM-Eval"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.6-Terra-Medium-Thinking",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "DeepSeek-V4-Pro",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Llama-3.3-70B-Instruct",
    "version": "3.3",
    "role": "system"
   },
   {
    "name": "Meta-Llama-3.1-8B-Instruct",
    "version": "3.1",
    "role": "system"
   },
   {
    "name": "Qwen3-Coder-480B",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "Qwen3-235B-A22B-Instruct-2507",
    "version": "2507",
    "role": "baseline"
   },
   {
    "name": "Qwen3-30B-A3B-Instruct-2507",
    "version": "2507",
    "role": "baseline"
   },
   {
    "name": "Qwen2.5-Coder-7B",
    "version": "2.5",
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-terra-medium",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": -72,
   "frontier_at_paper": "GPT-5.5 Pro (2026-04-23)",
   "flagship_releases_since": 0
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 47
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2605.31495",
  "date": "2026-05-29",
  "title": "General-purpose LLMs as Constrained Crystal Composition Generators",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4-nano",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4-nano",
   "newest_release": "2026-03-17",
   "newest_class": "frontier-api",
   "lag_days": 73,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 47
  },
  "subject_codes": {
   "sol": {
    "subject": "functional-solids",
    "question": "method-capability"
   },
   "opus": {
    "subject": "functional-solids",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.00065",
  "date": "2026-05-19",
  "title": "Beyond Text and Tables: Vision-Language Model Integration in ComProScanner for Extracting Materials Data from Scientific Figures with High Accuracy",
  "scope_group": "materials-science",
  "use": "literature-knowledge",
  "kind": "method",
  "harness": "off-the-shelf: CrewAI",
  "harness_class": "off-the-shelf",
  "base_models": 6,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated; model selection dated 15 April 2026",
  "code_urls": [
   "https://github.com/slimeslab/ComProScanner"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Gemini-3-Flash-Preview",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini-2.5-Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5-Chat-Latest",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.1",
    "version": null,
    "role": "system"
   },
   {
    "name": "claude-sonnet-4-6",
    "version": "claude-sonnet-4-6",
    "role": "system"
   },
   {
    "name": "deepseek/deepseek-chat",
    "version": "deepseek/deepseek-chat",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-4.6",
   "newest_release": "2026-02-17",
   "newest_class": "frontier-api",
   "lag_days": 91,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 1,
   "bibliography_size": 35
  },
  "subject_codes": {
   "sol": {
    "subject": "functional-solids",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.00315",
  "date": "2026-05-29",
  "title": "Coupling Language Models with Physics-based Simulation for Synthesis of Inorganic Materials",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "off-the-shelf: DSPy ReAct",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated; model reference accessed 2025-08-18",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT4o",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-4o",
   "newest_release": "2024-05-13",
   "newest_class": "frontier-api",
   "lag_days": 746,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 32
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 19
  },
  "subject_codes": {
   "sol": {
    "subject": "functional-solids",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2606.01089",
  "date": "2026-05-31",
  "title": "A literature-grounded scientific reasoning framework for defect-engineered TiO$_{2}$ photocatalysts",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-based",
    "version": null,
    "role": "system"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "catalysts",
    "question": "targeted-design"
   },
   "opus": {
    "subject": "catalysts",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.02418",
  "date": "2026-06-01",
  "title": "Evolutionary Discovery of Bivariate Bicycle Codes with LLM-Guided Search",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "application",
  "harness": "off-the-shelf: OpenEvolve",
  "harness_class": "off-the-shelf",
  "base_models": 6,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/qiskit-community/qcode-discovery"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Gemini 3 Flash Preview",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Opus 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "GPT-5.2",
    "version": "5.2",
    "role": "system"
   },
   {
    "name": "Gemini 3 Pro Preview",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.3-Codex",
    "version": "5.3-Codex",
    "role": "system"
   },
   {
    "name": "Gemini 3.1 Pro Preview",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gemini-3.1-pro",
   "newest_release": "2026-02-19",
   "newest_class": "frontier-api",
   "lag_days": 102,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 4,
   "bibliography_size": 134
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "targeted-design"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "targeted-design"
   }
  }
 },
 {
  "id": "2606.02434",
  "date": "2026-06-01",
  "title": "Bridging the Sim-to-Real Gap in Semiconductor Visual Program Synthesis via Input Binarization",
  "scope_group": "other",
  "use": "data-interpretation",
  "kind": "method",
  "harness": "off-the-shelf: LLaMA-Factory (linked release)",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/YusukeO/eusipco2026-sem-dsl"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen3-VL-8B",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "qwen3-vl-8b",
   "newest_release": "2025-10-15",
   "newest_class": "open-weights",
   "lag_days": 229,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 16
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 14
  },
  "subject_codes": {
   "sol": {
    "subject": "semiconductor-optical",
    "question": "method-capability"
   },
   "opus": {
    "subject": "semiconductor-optical",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.03300",
  "date": "2026-06-02",
  "title": "A proof of an identity for the critical exponents of jamming",
  "scope_group": "condensed-matter",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://doi.org/10.5281/zenodo.20478427"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Opus 4.7",
    "version": "4.7",
    "role": "system"
   },
   {
    "name": "Sonnet 4.6",
    "version": "4.6",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.7",
   "newest_release": "2026-04-16",
   "newest_class": "frontier-api",
   "lag_days": 47,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 24
  },
  "subject_codes": {
   "sol": {
    "subject": "molecular-soft",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "molecular-soft",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2606.03764",
  "date": "2026-06-02",
  "title": "RamanGPT: Bidirectional Mapping Between Crystal Structures and Raman Spectra with Graph Neural Networks and Generative Transformers",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/atomgptlab/atomgpt",
   "https://github.com/atomgptlab/alignn"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Mistral-7B-Instruct",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "mistral-7b",
   "newest_release": "2023-09-27",
   "newest_class": "open-weights",
   "lag_days": 979,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 32
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 59
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.05050",
  "date": "2026-06-03",
  "title": "Autonomous heterogeneous catalyst discovery with a self-evolving multi-agent digital twin",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/AI4QC/catdt"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "DeepSeek V4 Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.5",
    "version": "GPT-5.5",
    "role": "system"
   },
   {
    "name": "GPT-5.5",
    "version": "GPT-5.5",
    "role": "baseline"
   },
   {
    "name": "sentence-transformers/all-MiniLM-L6-v2",
    "version": "all-MiniLM-L6-v2",
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "deepseek-v4-pro",
   "newest_release": "2026-04-24",
   "newest_class": "open-weights",
   "lag_days": 40,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 2
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 77
  },
  "subject_codes": {
   "sol": {
    "subject": "catalysts",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "catalysts",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.05472",
  "date": "2026-06-03",
  "title": "Realistic quantum device data synthesized by consumer AI and how to identify it",
  "scope_group": "condensed-matter",
  "use": "data-interpretation",
  "kind": "application",
  "harness": "off-the-shelf: ChatGPT Data Analyst",
  "harness_class": "off-the-shelf",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "Initial tests 2023; manuscript written 2026; per-model dates not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "ChaptGPT 5.4",
    "version": "5.4",
    "role": "system"
   },
   {
    "name": "4o",
    "version": "4o",
    "role": "system"
   },
   {
    "name": "5.2",
    "version": "5.2",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4",
   "newest_release": "2026-03-05",
   "newest_class": "frontier-api",
   "lag_days": 90,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2606.06415",
  "date": "2026-06-04",
  "title": "PolyGraphPy: A unified Python framework for atomistic simulation and machine learning-driven polymer design",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/SASP-lab/PolyGraphPy"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "small GPT-2",
    "version": "124M; openai-community/gpt2; revision unspecified",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "small-gpt-2",
   "newest_release": "2019-02-14",
   "newest_class": "open-weights",
   "lag_days": 2667,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 32
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 77
  },
  "subject_codes": {
   "sol": {
    "subject": "molecular-soft",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "molecular-soft",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2606.07314",
  "date": "2026-06-05",
  "title": "QBugLM: An Agentic Benchmarking Framework for LLM-based Quantum Software Debugging",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "benchmark",
  "harness": "off-the-shelf: Strands Agents",
  "harness_class": "off-the-shelf",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated; March 2026 dates pricing only",
  "code_urls": [
   "https://github.com/qachub/qbuglm",
   "https://github.com/Qiskit/qiskit-aer"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Sonnet 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "Qwen3 Coder Next",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-4.6",
   "newest_release": "2026-02-17",
   "newest_class": "frontier-api",
   "lag_days": 108,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 3,
   "bibliography_size": 33
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2606.07570",
  "date": "2026-05-26",
  "title": "Can LLMs extract scientific consensus? A case study in high-temperature superconductivity",
  "scope_group": "condensed-matter",
  "use": "literature-knowledge",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 5,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "gpt-5-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "gpt-4.1-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "gpt-4.1",
    "version": null,
    "role": "system"
   },
   {
    "name": "o4-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "gpt-5.4",
    "version": null,
    "role": "system"
   },
   {
    "name": "BERT",
    "version": "MaartenGr/BERTopic_Wikipedia",
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4",
   "newest_release": "2026-03-05",
   "newest_class": "frontier-api",
   "lag_days": 82,
   "frontier_at_paper": "Grok 4.3 (2026-04-30)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 39
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-materials",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-materials",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2606.07712",
  "date": "2026-06-05",
  "title": "MatMind: A Structure-Activity Knowledge-Driven Generative Foundation Model for Materials Science",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "none",
  "harness_class": "none",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "S1-Base 8B",
    "version": null,
    "role": "system"
   },
   {
    "name": "MatMind",
    "version": null,
    "role": "system"
   },
   {
    "name": "LLM-Prop",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "MatBERT-109M",
    "version": null,
    "role": "baseline"
   }
  ],
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 84
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.09462",
  "date": "2026-06-08",
  "title": "Free fermions in disguise without exponential degeneracies",
  "scope_group": "condensed-matter",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "none",
  "harness_class": "none",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "ChatGPT 5.4",
    "version": "5.4",
    "role": "system"
   },
   {
    "name": "ChatGPT 5.5 Pro",
    "version": "5.5",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5-pro",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 46,
   "frontier_at_paper": "Claude Opus 4.8 (2026-05-28)",
   "flagship_releases_since": 2
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 18
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2606.12916",
  "date": "2026-06-11",
  "title": "MDForge: Agentic Molecular Dynamics Pipeline Design under Sparse Simulator Feedback",
  "scope_group": "other",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Zehong-Wang/MDForge"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Opus 4.7",
    "version": "4.7",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.7",
   "newest_release": "2026-04-16",
   "newest_class": "frontier-api",
   "lag_days": 56,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 119
  }
 },
 {
  "id": "2606.13380",
  "date": "2026-06-11",
  "title": "An LLM System for Autonomous Variational Quantum Circuit Design",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 4,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Qyusu/astronaut",
   "https://github.com/Qyusu/astronaut-artifact",
   "https://github.com/Qyusu/qxmt"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "o3-deep-research",
    "version": "o3-deep-research-2025-06-26",
    "role": "system"
   },
   {
    "name": "GPT-5",
    "version": "gpt-5-2025-08-07",
    "role": "system"
   },
   {
    "name": "gpt-5-mini-2025-08-07",
    "version": "gpt-5-mini-2025-08-07",
    "role": "system"
   },
   {
    "name": "gpt-5-nano-2025-08-07",
    "version": "gpt-5-nano-2025-08-07",
    "role": "system"
   },
   {
    "name": "text-embedding-3-small",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5-nano",
   "newest_release": "2025-08-07",
   "newest_class": "frontier-api",
   "lag_days": 308,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 18
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 71
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.13811",
  "date": "2026-06-11",
  "title": "Aligning Quantum Operators with Large Language Models",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Granite 4.0 Micro",
    "version": "Granite-4.0-micro",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "granite-4.0-micro",
   "newest_release": "2025-10-02",
   "newest_class": "open-weights",
   "lag_days": 252,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 17
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 25
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.17809",
  "date": "2026-06-16",
  "title": "Million-scale multimodal pollen microscopy with expert-guided foundation models",
  "scope_group": "other",
  "use": "data-interpretation",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 5,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/abiricz/pollen-ai-atlas-paper"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Qwen2.5-VL",
    "version": "Qwen2.5-VL-32B-AWQ",
    "role": "system"
   },
   {
    "name": "Qwen3-VL",
    "version": "Qwen3-VL-32B-Instruct-FP8",
    "role": "system"
   },
   {
    "name": "Qwen3.5",
    "version": "Qwen3.5-27B-FP8",
    "role": "system"
   },
   {
    "name": "Qwen3.6",
    "version": "Qwen3.6-27B-FP8",
    "role": "system"
   },
   {
    "name": "Gemma4",
    "version": "Gemma-4-31B-it",
    "role": "system"
   },
   {
    "name": "SBERT",
    "version": "all-MiniLM-L6-v2",
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gemma-4",
   "newest_release": "2026-04-02",
   "newest_class": "open-weights",
   "lag_days": 75,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 61
  }
 },
 {
  "id": "2606.18422",
  "date": "2026-06-16",
  "title": "Gatekeepers and Hallucinations: A Layered Evaluation Framework for LLM-Driven Quantum Circuit Generation",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "benchmark",
  "harness": "off-the-shelf: AWS Strands Agents SDK, custom platform",
  "harness_class": "off-the-shelf",
  "base_models": 8,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/UBC-CIC/Quantum-Matter-Institute-Streamlit-App"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Sonnet 4.5",
    "version": "4.5",
    "role": "system"
   },
   {
    "name": "Claude Opus 4.1",
    "version": "4.1",
    "role": "system"
   },
   {
    "name": "OpenAI OSS-120B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Nova Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen 3-32B",
    "version": null,
    "role": "system"
   },
   {
    "name": "DeepSeek R1",
    "version": null,
    "role": "system"
   },
   {
    "name": "Llama 4 Scout",
    "version": null,
    "role": "system"
   },
   {
    "name": "Llama 3 70B",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-4.5",
   "newest_release": "2025-09-29",
   "newest_class": "frontier-api",
   "lag_days": 260,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 17
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 22
  }
 },
 {
  "id": "2606.19152",
  "date": "2026-06-17",
  "title": "AdsMind: A Physics-Grounded Multi-Agent System for Self-Correcting Discovery of Adsorption Configurations on Heterogeneous Catalyst Surfaces",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 4,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/NagatoBigSeven/AdsMind",
   "https://github.com/hoon-ock/CatalystAIgent"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Gemini-2.5-Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude-Sonnet-4.6",
    "version": null,
    "role": "system"
   },
   {
    "name": "Grok-4",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4",
   "newest_release": "2026-03-05",
   "newest_class": "frontier-api",
   "lag_days": 104,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 60
  },
  "subject_codes": {
   "sol": {
    "subject": "catalysts",
    "question": "method-capability"
   },
   "opus": {
    "subject": "catalysts",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.20729",
  "date": "2026-06-17",
  "title": "LLM-Guided Test-Time Discovery of Quantum-Chemical Approximation Algorithms",
  "scope_group": "other",
  "use": "code-algorithm-generation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": true,
  "evaluation_dates": "accessed Sep. 2025 to Dec. 2025",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5",
    "version": null,
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5",
   "newest_release": "2025-08-07",
   "newest_class": "frontier-api",
   "lag_days": 314,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 18
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 35
  }
 },
 {
  "id": "2606.21395",
  "date": "2026-06-19",
  "title": "Atomistic Language Models Understand and Generate Materials",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 5,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen3-0.6B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-1.7B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-4B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-8B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-14B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen2.5-3B base",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-4o",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-4.1",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5.2",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-4o",
    "version": null,
    "role": "judge"
   },
   {
    "name": "gpt-4o-mini",
    "version": null,
    "role": "judge"
   },
   {
    "name": "CrystaLLM-large",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "CrystalTextLLM",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "PLAID++ Wyckoff",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "CrystalReasoner",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "CrysReas-Base (SFT only)",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "CrysReas-Thinking (SFT+CoT)",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "CrysReas-RL (SFT+GRPO)",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "MatterChat",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "MatBERT (CIF)",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "LLM-Prop (CIF)",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "CrysLLMGen-7B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "FlowLLM",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "unspecified LLM",
    "version": null,
    "role": "data-generation"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.2",
   "newest_release": "2025-12-11",
   "newest_class": "frontier-api",
   "lag_days": 190,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 10
  },
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 0,
   "bibliography_size": 75
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.21974",
  "date": "2026-06-20",
  "title": "Fine-Tuning Large Language Models for Quantum Reasoning",
  "scope_group": "adjacent-quantum",
  "use": "property-prediction",
  "kind": "method",
  "harness": "none",
  "harness_class": "none",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen3-8B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-8B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-OSS-120B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "DeepSeek-R1",
    "version": null,
    "role": "data-generation"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-oss-120b",
   "newest_release": "2025-08-05",
   "newest_class": "open-weights",
   "lag_days": 319,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 19
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 53
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.22375",
  "date": "2026-06-21",
  "title": "ARIA: A Causal-Aware Framework for Rescuing LLM Reasoning in Trustworthy Materials Discovery",
  "scope_group": "materials-science",
  "use": "property-prediction",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/yicao-elina/ARIA"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "DeepSeek-R1:8B",
    "version": null,
    "role": "system"
   },
   {
    "name": "DeepSeek-R1:8B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "DeepSeek-R1-8B",
    "version": null,
    "role": "judge"
   },
   {
    "name": "DeepSeek-R1:14B",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Qwen2.5:7B",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "GLM-5 (cloud)",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Qwen-3.5 (cloud)",
    "version": null,
    "role": "judge"
   },
   {
    "name": "all-MiniLM-L6-v2",
    "version": null,
    "role": "other"
   },
   {
    "name": "Self-RAG",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "glm-5",
   "newest_release": "2026-02-11",
   "newest_class": "open-weights",
   "lag_days": 130,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 9
  },
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 0,
   "bibliography_size": 41
  },
  "subject_codes": {
   "sol": {
    "subject": "semiconductor-optical",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.22376",
  "date": "2026-06-21",
  "title": "Vibe Calibration: Autonomous Bring-up of a 112-Qubit Superconducting Quantum Processor by a Skill-Orchestrating Language Agent",
  "scope_group": "adjacent-quantum",
  "use": "experimental-control",
  "kind": "application",
  "harness": "off-the-shelf: Claude Code v2.1.140; KimiCode; custom serving middleware",
  "harness_class": "off-the-shelf",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": true,
  "evaluation_dates": "Appendix A logs: 2026-05-10; other evaluations not dated",
  "code_urls": [
   "https://github.com/NVIDIA/Quantum-Calibration-Agent-Blueprint"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen3.6-35B-A3B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3.5-4B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Kimi K2.6",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "DeepSeek-V4",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "unspecified teacher LLM",
    "version": null,
    "role": "data-generation"
   }
  ],
  "model_lag": {
   "newest_model": "kimi-k2.6",
   "newest_release": "2026-04-20",
   "newest_class": "open-weights",
   "lag_days": 62,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 2,
   "bibliography_size": 90
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2606.22497",
  "date": "2026-06-21",
  "title": "Benchmarking Vision-Language Models for Microscopic Plant Image Understanding",
  "scope_group": "other",
  "use": "scientific-qa",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/tqwei05/PlantMicro"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-4o mini",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5 mini",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini 2.5-Flash-lite",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini 2.5-Flash",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini 2.5-Pro",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "CLIP-ViT/32",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "CLIP-ViT/16",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "LLaVA-7B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "LLaVA-13B",
    "version": null,
    "role": "system"
   },
   {
    "name": "LLaVA-Next-7B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "LLaVA-Next-13B",
    "version": null,
    "role": "system"
   },
   {
    "name": "QwenVLM-7B",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5-mini",
   "newest_release": "2025-08-07",
   "newest_class": "frontier-api",
   "lag_days": 318,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 18
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 68
  }
 },
 {
  "id": "2606.23571",
  "date": "2026-06-22",
  "title": "INCARBench: A Benchmark for Scientific Configuration in VASP INCAR by Large Language Models",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 15,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/liuzhiyangnku/incarbench"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.4 (Default)",
    "version": "5.4",
    "role": "system"
   },
   {
    "name": "GPT-5.4 (Extended Reasoning)",
    "version": "5.4",
    "role": "system"
   },
   {
    "name": "GPT-5.4 (No Reasoning)",
    "version": "5.4",
    "role": "system"
   },
   {
    "name": "GPT-5.3 (Codex)",
    "version": "5.3",
    "role": "system"
   },
   {
    "name": "Claude Opus 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "Claude Sonnet 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "Claude Haiku 4.5",
    "version": "4.5",
    "role": "system"
   },
   {
    "name": "Gemini 3.1 Pro Preview",
    "version": "3.1 Pro Preview",
    "role": "system"
   },
   {
    "name": "Gemini 2.5 Pro",
    "version": "2.5 Pro",
    "role": "system"
   },
   {
    "name": "DeepSeek-R1",
    "version": null,
    "role": "system"
   },
   {
    "name": "DeepSeek-V3",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3.6+ (Thinking)",
    "version": "3.6+",
    "role": "system"
   },
   {
    "name": "Qwen3.6+ (No-Thinking)",
    "version": "3.6+",
    "role": "system"
   },
   {
    "name": "Qwen3-Coder-Plus",
    "version": null,
    "role": "system"
   },
   {
    "name": "GLM-5.1",
    "version": "5.1",
    "role": "system"
   },
   {
    "name": "Kimi K2.5 (Thinking)",
    "version": "K2.5",
    "role": "system"
   },
   {
    "name": "Kimi K2.5 (No-Thinking)",
    "version": "K2.5",
    "role": "system"
   },
   {
    "name": "MiniMax-M2.7",
    "version": "M2.7",
    "role": "system"
   },
   {
    "name": "Llama 4 Maverick",
    "version": "4 Maverick",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "glm-5.1",
   "newest_release": "2026-04-07",
   "newest_class": "open-weights",
   "lag_days": 76,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 2,
   "bibliography_size": 39
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2606.24480",
  "date": "2026-06-23",
  "title": "Breaking Bottlenecks in Solid Electrolyte Discovery with Large Artificial Intelligence Models",
  "scope_group": "materials-science",
  "use": "research-orchestration",
  "kind": "review-perspective",
  "harness": "none",
  "harness_class": "none",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "BatteryBERT",
    "version": null,
    "role": "other"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "battery-materials",
    "question": "ai-workflow-perspective"
   },
   "opus": {
    "subject": "battery-materials",
    "question": "ai-workflow-perspective"
   }
  }
 },
 {
  "id": "2606.24808",
  "date": "2026-06-23",
  "title": "Large-Language-Model Discovery of Quantum LDPC Codes through Structured Concept Evolution",
  "scope_group": "adjacent-quantum",
  "use": "design-optimization",
  "kind": "method",
  "harness": "off-the-shelf: OpenEvolve",
  "harness_class": "off-the-shelf",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/algorithmicsuperintelligence/openevolve"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.4-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4-nano",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4-nano",
   "newest_release": "2026-03-17",
   "newest_class": "frontier-api",
   "lag_days": 98,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 3,
   "bibliography_size": 63
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.25870",
  "date": "2026-06-24",
  "title": "Evolving Quantum Error-Correcting Encodings for Molecular Simulation",
  "scope_group": "adjacent-quantum",
  "use": "design-optimization",
  "kind": "application",
  "harness": "off-the-shelf: AlphaEvolve Cloud API; ShinkaEvolve",
  "harness_class": "off-the-shelf",
  "base_models": 4,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/SakanaAI/ShinkaEvolve"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Gemini 3.1 Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini 3 Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 62,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 0,
   "bibliography_size": 32
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "targeted-design"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "targeted-design"
   }
  }
 },
 {
  "id": "2606.27231",
  "date": "2026-06-25",
  "title": "A hardware-safety-gated system for LLM-written native ARTIQ control code on a trapped-ion platform",
  "scope_group": "adjacent-quantum",
  "use": "experimental-control",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "Paper: not stated; supplement: June 2026 campaign, September 2026 rechecks/reclassification",
  "code_urls": [
   "https://gitlab.oit.duke.edu/dw346/artiq_mcp"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Opus 4.8",
    "version": "4.8",
    "role": "system"
   },
   {
    "name": "language-model generator (unspecified)",
    "version": null,
    "role": "data-generation"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.8",
   "newest_release": "2026-05-28",
   "newest_class": "frontier-api",
   "lag_days": 28,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 3,
   "bibliography_size": 36
  }
 },
 {
  "id": "2606.29459",
  "date": "2026-06-28",
  "title": "Interpretable Inverse Design of Metal-Organic Frameworks with Large Language Model Agents",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 6,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/kn1218/LLM4MOF"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "gpt-5.2",
    "version": "gpt-5.2",
    "role": "system"
   },
   {
    "name": "gpt-5.6-sol",
    "version": "gpt-5.6-sol",
    "role": "system"
   },
   {
    "name": "claude-opus-5",
    "version": "claude-opus-5",
    "role": "system"
   },
   {
    "name": "claude-sonnet-5",
    "version": "claude-sonnet-5",
    "role": "system"
   },
   {
    "name": "claude-haiku-4.5",
    "version": "claude-haiku-4.5",
    "role": "system"
   },
   {
    "name": "gpt-5.4-nano",
    "version": "gpt-5.4-nano",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-5",
   "newest_release": "2026-07-24",
   "newest_class": "frontier-api",
   "lag_days": -26,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 0
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "porous-frameworks",
    "question": "method-capability"
   },
   "opus": {
    "subject": "porous-frameworks",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.29592",
  "date": "2026-06-28",
  "title": "STEMGym: Benchmarking Sequential Decision-Making under Dose Budgets in Autonomous Electron Microscopy",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "benchmark",
  "harness": "custom; LangGraph planner; OpenRouter VLM calls",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/KurbanIntelligenceLab/STEMGym"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Haiku 4.5",
    "version": "anthropic/claude-haiku-4.5",
    "role": "baseline"
   },
   {
    "name": "GPT-5 mini",
    "version": "openai/gpt-5-mini",
    "role": "baseline"
   },
   {
    "name": "Llama 4 Scout",
    "version": "meta-llama/llama-4-scout",
    "role": "baseline"
   },
   {
    "name": "Gemini 2.0 Flash",
    "version": "google/gemini-2.0-flash-001",
    "role": "baseline"
   },
   {
    "name": "Claude Opus 4.7",
    "version": "anthropic/claude-opus-4.7",
    "role": "baseline"
   },
   {
    "name": "Gemini 3 Flash Preview",
    "version": "google/gemini-3-flash-preview",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.7",
   "newest_release": "2026-04-16",
   "newest_class": "frontier-api",
   "lag_days": 73,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 47
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2606.29667",
  "date": "2026-06-29",
  "title": "MatMMExtract: An Open-Source Pipeline for Panel-Level Extraction of Grounded Image-Text Pairs from Materials Science Literature",
  "scope_group": "materials-science",
  "use": "literature-knowledge",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 7,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated; pricing references accessed 2026-06-21",
  "code_urls": [
   "https://github.com/CMEG-IITR/matmmextract",
   "https://github.com/CMEG-IITR/cmpfig"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Gemini 3.5 Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "DeepSeek V4 Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini 3.1 Flash Lite",
    "version": null,
    "role": "system"
   },
   {
    "name": "Mistral Large 3",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4 Mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4 Nano",
    "version": null,
    "role": "system"
   },
   {
    "name": "MatSciBERT",
    "version": null,
    "role": "system"
   },
   {
    "name": "CLIP ViT-B/32",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Claude Sonnet 4.6",
    "version": null,
    "role": "writing"
   }
  ],
  "model_lag": {
   "newest_model": "gemini-3.5-flash",
   "newest_release": "2026-05-19",
   "newest_class": "frontier-api",
   "lag_days": 41,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 2
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 53
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "general",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2606.29687",
  "date": "2026-06-29",
  "title": "A Machine-Verified Proof of a Quantum-Optimization Conjecture",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "custom: LeanScriber",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/urikol/QuantumOptimization",
   "https://github.com/urikol/leanscriber"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Fable 5",
    "version": "5",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-fable-5",
   "newest_release": "2026-06-09",
   "newest_class": "frontier-api",
   "lag_days": 20,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 0
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 3,
   "bibliography_size": 47
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2606.29717",
  "date": "2026-06-29",
  "title": "Optimizing Expert-Designed Crystal Graph Networks for Band-Gap Prediction with an Autonomous LLM Research Loop",
  "scope_group": "materials-science",
  "use": "code-algorithm-generation",
  "kind": "application",
  "harness": "off-the-shelf: Claude Code",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/karpathy/autoresearch"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Opus 4.8",
    "version": "4.8",
    "role": "system"
   },
   {
    "name": "MatInFormer",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.8",
   "newest_release": "2026-05-28",
   "newest_class": "frontier-api",
   "lag_days": 32,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 36
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2606.29966",
  "date": "2026-06-29",
  "title": "RiverONE: Generating Knowledge-Intensive VLM by Simulated Quantum Machines",
  "scope_group": "adjacent-quantum",
  "use": "data-interpretation",
  "kind": "method",
  "harness": "none",
  "harness_class": "none",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/THeWakeSystems/RiverOne",
   "https://github.com/CharlesCNorton/qiskit-calibration-drift"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "InternVL3_5-4B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Opus 4.6",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GLM-5V-Turbo",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "NVIDIA Ising Calibration 1",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.5-122B-A10B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.5-122B-A10B",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "InternVL3_5-2B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "InternVL3_5-4B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "InternVL3_5-8B",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "glm-5v",
   "newest_release": "2026-04-01",
   "newest_class": "frontier-api",
   "lag_days": 89,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 38
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.00924",
  "date": "2026-07-01",
  "title": "Graph-Native Reinforcement Learning Enables Traceable Scientific Hypothesis Generation through Conceptual Recombination",
  "scope_group": "materials-science",
  "use": "reasoning-proof",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/lamm-mit/graph-preflexor-grpo",
   "https://github.com/datalab-to/marker"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Graph-PRefLexOR-8B",
    "version": "Graph-Preflexor-8b_12292025",
    "role": "system"
   },
   {
    "name": "Graph-PRefLexOR-3B",
    "version": "Graph-Preflexor-3b_08012026",
    "role": "system"
   },
   {
    "name": "Graph-PRefLexOR-1.7B",
    "version": "Graph-Preflexor-1.7b_08012026",
    "role": "system"
   },
   {
    "name": "Qwen3-8B",
    "version": "Qwen3-8B",
    "role": "baseline"
   },
   {
    "name": "Llama-3.2-3B-Instruct",
    "version": "Llama-3.2-3B-Instruct",
    "role": "baseline"
   },
   {
    "name": "Qwen3-1.7B",
    "version": "Qwen3-1.7B",
    "role": "baseline"
   },
   {
    "name": "GPT-5.1",
    "version": "GPT-5.1",
    "role": "data-generation"
   },
   {
    "name": "GPT-5-nano",
    "version": "GPT-5-nano",
    "role": "data-generation"
   },
   {
    "name": "gpt-4o-mini",
    "version": "gpt-4o-mini",
    "role": "data-generation"
   },
   {
    "name": "gpt-5.4",
    "version": "gpt-5.4",
    "role": "data-generation"
   },
   {
    "name": "Claude opus-4.7",
    "version": "Claude opus-4.7",
    "role": "judge"
   },
   {
    "name": "grok-4-1-fast-non-reasoning",
    "version": "grok-4-1-fast-non-reasoning",
    "role": "judge"
   },
   {
    "name": "google/embeddinggemma_300m",
    "version": null,
    "role": "other"
   },
   {
    "name": "all-MiniLM-L6-v2",
    "version": null,
    "role": "other"
   },
   {
    "name": "BAAI/bge-base-en-v1.5",
    "version": null,
    "role": "other"
   },
   {
    "name": "questioner model",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.7",
   "newest_release": "2026-04-16",
   "newest_class": "frontier-api",
   "lag_days": 76,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 57
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.00939",
  "date": "2026-07-01",
  "title": "Leveraging LLM-Based Agentic Systems to Generate Quantum Applications for Test Optimization",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://doi.org/10.5281/zenodo.21094837",
   "https://doi.org/10.5281/zenodo.21094908"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "claude-opus-4-7",
    "version": "claude-opus-4-7",
    "role": "system"
   },
   {
    "name": "DeepSeek-V4-Flash",
    "version": "DeepSeek-V4-Flash",
    "role": "system"
   },
   {
    "name": "llama4-maverick-17b",
    "version": "llama4-maverick-17b",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "deepseek-v4-flash",
   "newest_release": "2026-04-24",
   "newest_class": "open-weights",
   "lag_days": 68,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 42
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.02329",
  "date": "2026-07-02",
  "title": "Grounded autonomous research: a fault-tolerant LLM pipeline from corpus to manuscript in frontier computational physics",
  "scope_group": "condensed-matter",
  "use": "research-orchestration",
  "kind": "application",
  "harness": "off-the-shelf: Claude Code CLI",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": true,
  "evaluation_dates": "April–May 2026",
  "code_urls": [
   "https://doi.org/10.5281/zenodo.21126996"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Opus",
    "version": null,
    "role": "system"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 38
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-materials",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-materials",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.02934",
  "date": "2026-07-03",
  "title": "MatPhaseBench: A Semantics-Guided Benchmark for Materials Phase Diagrams Understanding",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "benchmark",
  "harness": "none",
  "harness_class": "none",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Davidwhw/MatPhaseBench"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude-Opus-4.8",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GLM-5V-Turbo",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini-3.1-Pro",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.6-Plus",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.6-Flash",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini-3.1-Flash-Lite",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.6-27B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.6-35B-A3B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GLM-4.6V",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GLM-4.6V-Flash",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Llama-4-Maverick",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Llama-4-Scout",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.6-Plus",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "GLM-5.1",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "XLNet",
    "version": "xlnet-large-cased",
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.8",
   "newest_release": "2026-05-28",
   "newest_class": "frontier-api",
   "lag_days": 36,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 43
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2607.03105",
  "date": "2026-07-03",
  "title": "ORBIT-Q: Dual-axis benchmarking of autonomous agents in scientific quantum programming",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "benchmark",
  "harness": "off-the-shelf: Harbor; customized adapters invoking Codex or Claude Code",
  "harness_class": "off-the-shelf",
  "base_models": 4,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/sxzgroup/ORBIT-Q"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Claude Opus-4.8",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Sonnet-4.6",
    "version": null,
    "role": "system"
   },
   {
    "name": "GLM-5.2",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "glm-5.2",
   "newest_release": "2026-06-16",
   "newest_class": "open-weights",
   "lag_days": 17,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 0
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 47
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2607.03193",
  "date": "2026-07-03",
  "title": "Self-Specializing Vision-Language Transmon Chip Calibration in a Physics-Grounded Environment",
  "scope_group": "adjacent-quantum",
  "use": "experimental-control",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "gemini-3.1-pro-preview",
    "version": "gemini-3.1-pro-preview",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gemini-3.1-pro",
   "newest_release": "2026-02-19",
   "newest_class": "frontier-api",
   "lag_days": 134,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 0,
   "bibliography_size": 30
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.04072",
  "date": "2026-07-05",
  "title": "Benchmarking API Drift in LLM-Generated Quantum Code Across Successive SDK Versions",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "benchmark",
  "harness": "custom; off-the-shelf: Codex CLI",
  "harness_class": "off-the-shelf",
  "base_models": 16,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/arasyi/quantum-api-drift"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.4 (Codex CLI)",
    "version": "gpt-5.4-medium",
    "role": "system"
   },
   {
    "name": "Claude Opus 4.7",
    "version": "claude-opus-4-7",
    "role": "system"
   },
   {
    "name": "GPT-5.3-codex",
    "version": "gpt-5.3-codex",
    "role": "system"
   },
   {
    "name": "GPT-5.4 (API)",
    "version": "gpt-5.4",
    "role": "system"
   },
   {
    "name": "Grok 4.20",
    "version": "grok-4-20",
    "role": "system"
   },
   {
    "name": "Claude Sonnet 4.6",
    "version": "claude-sonnet-4-6",
    "role": "system"
   },
   {
    "name": "GPT-5.4 Nano",
    "version": "gpt-5.4-nano",
    "role": "system"
   },
   {
    "name": "Kimi K2",
    "version": "kimi-k2-instruct",
    "role": "system"
   },
   {
    "name": "Nemotron 3 Super",
    "version": "nemotron-3-super-120b",
    "role": "system"
   },
   {
    "name": "GPT-OSS 120B",
    "version": "gpt-oss-120b",
    "role": "system"
   },
   {
    "name": "Gemma 4 31B",
    "version": "gemma-4-31b-it",
    "role": "system"
   },
   {
    "name": "Qwen3 Coder",
    "version": "qwen3-coder",
    "role": "system"
   },
   {
    "name": "Devstral 2",
    "version": "devstral-2-123b",
    "role": "system"
   },
   {
    "name": "Qwen3 32B",
    "version": "qwen3-32b",
    "role": "system"
   },
   {
    "name": "GLM 4.7 Flash",
    "version": "glm-4.7-flash",
    "role": "system"
   },
   {
    "name": "Llama 3.3",
    "version": "llama-3.3-70b-instruct",
    "role": "system"
   },
   {
    "name": "Llama 4 Scout",
    "version": "llama-4-scout-17b-16e",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.7",
   "newest_release": "2026-04-16",
   "newest_class": "frontier-api",
   "lag_days": 80,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 25
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2607.05805",
  "date": "2026-07-07",
  "title": "Onnes: A Physics-Grounded Multi-Agent LLM Simulator for Cryogenic Fault Diagnosis in Quantum Computing Infrastructure",
  "scope_group": "adjacent-quantum",
  "use": "data-interpretation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Onnes-Research/onnes"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Opus 4.8",
    "version": null,
    "role": "system"
   },
   {
    "name": "Google Gemini 3.1 Pro",
    "version": "gemini-3.1-pro-preview",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.8",
   "newest_release": "2026-05-28",
   "newest_class": "frontier-api",
   "lag_days": 40,
   "frontier_at_paper": "Claude Fable 5 (2026-06-09)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 31
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2607.07554",
  "date": "2026-07-08",
  "title": "RubriQ: Rubric-Guided Group Relative Policy Optimization for Constraint-Aware Quantum Circuit Synthesis",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://huggingface.co/spaces/qcsyn/rubriq-demo"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen2.5-Coder-7B",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "qwen2.5-coder-7b",
   "newest_release": "2024-09-19",
   "newest_class": "open-weights",
   "lag_days": 657,
   "frontier_at_paper": "Grok 4.5 (2026-07-08)",
   "flagship_releases_since": 34
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 65
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.07597",
  "date": "2026-07-08",
  "title": "Quantum Software Engineering in Practice: FPGA and AI Integration for Quantum Certification",
  "scope_group": "adjacent-quantum",
  "use": "design-optimization",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/jose-manuel-suarez/qacccert"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Mistral-7B-Instruct-v0.2",
    "version": "v0.2",
    "role": "system"
   },
   {
    "name": "DistilGPT2",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "mistral-7b-v0.2",
   "newest_release": "2023-12-11",
   "newest_class": "open-weights",
   "lag_days": 940,
   "frontier_at_paper": "Grok 4.5 (2026-07-08)",
   "flagship_releases_since": 34
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 25
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.07604",
  "date": "2026-07-08",
  "title": "Human and LLM Collaboration for Accelerated Materials Synthesis and Discovery",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "application",
  "harness": "none",
  "harness_class": "none",
  "base_models": 7,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "o1",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT5 thinking / gpt-5-chat",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT 5.2 extended thinking",
    "version": null,
    "role": "system"
   },
   {
    "name": "4o",
    "version": null,
    "role": "system"
   },
   {
    "name": "claude-haiku-4-5",
    "version": null,
    "role": "system"
   },
   {
    "name": "claude-opus-4-5",
    "version": null,
    "role": "system"
   },
   {
    "name": "claude-sonnet-4-5",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.2",
   "newest_release": "2025-12-11",
   "newest_class": "frontier-api",
   "lag_days": 209,
   "frontier_at_paper": "Grok 4.5 (2026-07-08)",
   "flagship_releases_since": 11
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 42
  },
  "subject_codes": {
   "sol": {
    "subject": "functional-solids",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "functional-solids",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2607.07857",
  "date": "2026-07-08",
  "title": "Multi-agent Autoformalization of Tensor Network Theory",
  "scope_group": "condensed-matter",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 17,
  "dated_system_snapshot": false,
  "eval_dates_stated": true,
  "evaluation_dates": "6 February–29 June 2026 (development logs)",
  "code_urls": [
   "https://github.com/LionSR/TNLean",
   "https://github.com/math-xmum/Brouwer",
   "https://github.com/LionSR/Brouwer"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Opus 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "GPT-5.5",
    "version": "5.5",
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": "5.4",
    "role": "system"
   },
   {
    "name": "Claude Opus 4.7",
    "version": "4.7",
    "role": "system"
   },
   {
    "name": "GPT-5.2",
    "version": "5.2",
    "role": "system"
   },
   {
    "name": "Claude Sonnet 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "Claude Opus 4.8",
    "version": "4.8",
    "role": "system"
   },
   {
    "name": "DeepSeek-pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini 3.1 Pro",
    "version": "3.1",
    "role": "system"
   },
   {
    "name": "GPT-5.4-mini",
    "version": "5.4-mini",
    "role": "system"
   },
   {
    "name": "Other models (7)",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.8",
   "newest_release": "2026-05-28",
   "newest_class": "frontier-api",
   "lag_days": 41,
   "frontier_at_paper": "Grok 4.5 (2026-07-08)",
   "flagship_releases_since": 2
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 67
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.09007",
  "date": "2026-07-10",
  "title": "Benchmarking Large Language Models on Repairing Qiskit Programs using Bugs4Q",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 4,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/slrrla/APR-new-experiment/tree/APRFinal",
   "https://github.com/Z-928/Bugs4Q-Framework",
   "https://github.com/Z-928/Bugs4Q-NA"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-4o-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5o-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Opus",
    "version": null,
    "role": "data-generation"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4-mini",
   "newest_release": "2026-03-17",
   "newest_class": "frontier-api",
   "lag_days": 115,
   "frontier_at_paper": "GPT-5.6 Sol Pro (2026-07-09)",
   "flagship_releases_since": 7
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 14
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2607.09195",
  "date": "2026-07-10",
  "title": "Toward Auditable AI Scientists: A Hypothesis Evolution Protocol for LLM Agents",
  "scope_group": "materials-science",
  "use": "research-orchestration",
  "kind": "method",
  "harness": "custom; OpenAI Agents SDK",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "judge"
   },
   {
    "name": "GPT-5.4-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-4.1",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 78,
   "frontier_at_paper": "GPT-5.6 Sol Pro (2026-07-09)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 0,
   "bibliography_size": 46
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.09789",
  "date": "2026-07-08",
  "title": "PHITSBench: an execution-scored benchmark for AI-assisted PHITS radiation-transport input generation using natural language",
  "scope_group": "other",
  "use": "simulation-workflow",
  "kind": "benchmark",
  "harness": "custom; off-the-shelf: Codex CLI v0.125.0; OpenAI Agents SDK v0.8.4",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated; April–May 2026 describes model release",
  "code_urls": [
   "https://github.com/openai/codex",
   "https://github.com/openai/openai-agents-python"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "unnamed LLM",
    "version": null,
    "role": "data-generation"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4",
   "newest_release": "2026-03-05",
   "newest_class": "frontier-api",
   "lag_days": 125,
   "frontier_at_paper": "Grok 4.5 (2026-07-08)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 46
  }
 },
 {
  "id": "2607.10057",
  "date": "2026-07-11",
  "title": "Quantum Circuit Vision: Cost-Aware Evaluation of Visual AI Agents for Quantum Code Generation",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "benchmark",
  "harness": "off-the-shelf: kiro-cli",
  "harness_class": "off-the-shelf",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Opus 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "Sonnet 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "Haiku 4.5",
    "version": "4.5",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-4.6",
   "newest_release": "2026-02-17",
   "newest_class": "frontier-api",
   "lag_days": 144,
   "frontier_at_paper": "GPT-5.6 Sol Pro (2026-07-09)",
   "flagship_releases_since": 9
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 48
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2607.14082",
  "date": "2026-07-15",
  "title": "Building Shor's Algorithm in Lean: An Agentic Formalization of Quantum Attacks on RSA-2048 and P-256",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/QudeLeap/Lean-QuantumAlg"
  ],
  "ai_use_statement": false,
  "models": [],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 2,
   "bibliography_size": 34
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.15535",
  "date": "2026-07-17",
  "title": "Symbolic Predicate-Guided Language Agents for Inverse Design of Perovskite Oxides",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/ahrehd0506/Catalyst-Design-Agent"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5-mini",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5-mini",
   "newest_release": "2025-08-07",
   "newest_class": "frontier-api",
   "lag_days": 344,
   "frontier_at_paper": "GPT-5.6 Sol Pro (2026-07-09)",
   "flagship_releases_since": 20
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "functional-solids",
    "question": "method-capability"
   },
   "opus": {
    "subject": "functional-solids",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.17012",
  "date": "2026-07-19",
  "title": "Schema-Bound LLM Control of Scientific Instrumentation through Model Context Protocol Skills",
  "scope_group": "other",
  "use": "experimental-control",
  "kind": "method",
  "harness": "off-the-shelf: LangChain ReAct",
  "harness_class": "off-the-shelf",
  "base_models": 5,
  "dated_system_snapshot": false,
  "eval_dates_stated": true,
  "evaluation_dates": "May 2026 (five-model probe)",
  "code_urls": [
   "https://github.com/NUANCE-IT/nuance-mcp"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "llama3.2",
    "version": null,
    "role": "system"
   },
   {
    "name": "qwen2.5",
    "version": "qwen2.5:7b",
    "role": "system"
   },
   {
    "name": "qwen2.5-coder",
    "version": null,
    "role": "system"
   },
   {
    "name": "nemotron",
    "version": null,
    "role": "system"
   },
   {
    "name": "mistral",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "llama-3.2",
   "newest_release": "2024-09-25",
   "newest_class": "open-weights",
   "lag_days": 662,
   "frontier_at_paper": "GPT-5.6 Sol Pro (2026-07-09)",
   "flagship_releases_since": 35
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 42
  }
 },
 {
  "id": "2607.17417",
  "date": "2026-07-19",
  "title": "Grounded verification of chemical and materials reasoning: detection is the bottleneck",
  "scope_group": "materials-science",
  "use": "scientific-qa",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 6,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/KurbanIntelligenceLab/grounded-matsci"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Sonnet 5",
    "version": "anthropic/claude-sonnet-5",
    "role": "system"
   },
   {
    "name": "GPT-5.5",
    "version": "openai/gpt-5.5",
    "role": "system"
   },
   {
    "name": "DeepSeek v4 Pro",
    "version": "deepseek/deepseek-v4-pro",
    "role": "system"
   },
   {
    "name": "Qwen3-235B",
    "version": "qwen/qwen3-235b-a22b-2507",
    "role": "system"
   },
   {
    "name": "Claude Haiku 4.5",
    "version": "anthropic/claude-haiku-4.5",
    "role": "system"
   },
   {
    "name": "GPT-5.4 mini",
    "version": "openai/gpt-5.4-mini",
    "role": "system"
   },
   {
    "name": "frontier reasoning-class LLM-as-judge",
    "version": null,
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-5",
   "newest_release": "2026-06-30",
   "newest_class": "frontier-api",
   "lag_days": 19,
   "frontier_at_paper": "GPT-5.6 Sol Pro (2026-07-09)",
   "flagship_releases_since": 2
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 34
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.17498",
  "date": "2026-07-20",
  "title": "LLM-Driven Cross-Paradigm Design for Quantum Optimal Control",
  "scope_group": "adjacent-quantum",
  "use": "design-optimization",
  "kind": "method",
  "harness": "off-the-shelf: OpenAI Codex",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 88,
   "frontier_at_paper": "GPT-5.6 Sol Pro (2026-07-09)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 47
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.19198",
  "date": "2026-07-21",
  "title": "ATLAS: A Foundation Neural Sampler for Amorphous Materials",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "LLM",
    "version": null,
    "role": "system"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 73
  },
  "subject_codes": {
   "sol": {
    "subject": "alloys",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.19506",
  "date": "2026-07-21",
  "title": "Hybrid LLM-Guided Search for Quantum Reservoir Architecture Design",
  "scope_group": "adjacent-quantum",
  "use": "design-optimization",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "LLM",
    "version": null,
    "role": "system"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 13
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.20058",
  "date": "2026-07-22",
  "title": "Reading and Steering Representations of Materials-Science Mechanisms in an Open-Weight Language Model",
  "scope_group": "materials-science",
  "use": "scientific-qa",
  "kind": "other",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": true,
  "evaluation_dates": "July–September 2026; SI S1 records July 14–September 2 milestones; judge date unspecified",
  "code_urls": [
   "https://github.com/lamm-mit/Substrates"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "google/gemma-4-E4B-it",
    "version": "a4c2d58be94dda072b918d9db64ee85c8ed34e3f",
    "role": "system"
   },
   {
    "name": "google/gemma-4-12B-it",
    "version": "707f0a3b8a3c7ad586ed01e27eafbad8a27dd0f7",
    "role": "system"
   },
   {
    "name": "google/gemma-4-31B-it",
    "version": "842da3794eaa0b77d5f08bae87a17459d91ff475",
    "role": "system"
   },
   {
    "name": "gpt-5.5",
    "version": null,
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "gemma-4-12b",
   "newest_release": "2026-06-03",
   "newest_class": "open-weights",
   "lag_days": 49,
   "frontier_at_paper": "GPT-5.6 Sol Pro (2026-07-09)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 52
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2607.20225",
  "date": "2026-07-22",
  "title": "DQAOA-GPT: AI-Accelerated Distributed Quantum Optimization for Combinatorial Problems",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://developer.nvidia.com/cuda-q"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT",
    "version": null,
    "role": "system"
   },
   {
    "name": "ChatGPT (OpenAI)",
    "version": null,
    "role": "writing"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 27
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.21533",
  "date": "2026-07-23",
  "title": "Benchmarking Agents for Proving Theorems in Quantum Algorithms and Quantum Information",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "benchmark",
  "harness": "off-the-shelf: Codex; Claude Code",
  "harness_class": "off-the-shelf",
  "base_models": 4,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated; pricing fixed 2026-07-20 and 2026-07-22",
  "code_urls": [
   "https://github.com/QudeLeap/Lean-QuantumAlg-Bench",
   "https://github.com/QuAIR/Lean-QIT-Bench"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.5",
    "version": "gpt-5.5",
    "role": "system"
   },
   {
    "name": "Kimi K3",
    "version": "k3[1m]",
    "role": "system"
   },
   {
    "name": "DeepSeek V4-Pro",
    "version": "deepseek-v4-pro[1m]",
    "role": "system"
   },
   {
    "name": "MiniMax M3",
    "version": "MiniMax M3",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "kimi-k3",
   "newest_release": "2026-07-16",
   "newest_class": "open-weights",
   "lag_days": 7,
   "frontier_at_paper": "GPT-5.6 Sol Pro (2026-07-09)",
   "flagship_releases_since": 0
  },
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 1,
   "bibliography_size": 55
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2607.21811",
  "date": "2026-07-23",
  "title": "Efficient Unclonable Encryption from Pauli Eigenstates",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "other",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/sragavan99/lean-unclonable-enc-paulis"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT-5.6 Sol Ultra",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.6 Sol Ultra",
    "version": null,
    "role": "writing"
   },
   {
    "name": "Codex GPT-5.6 Sol",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-sol-ultra",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 14,
   "frontier_at_paper": "GPT-5.6 Sol Pro (2026-07-09)",
   "flagship_releases_since": 0
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 19
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "targeted-design"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "targeted-design"
   }
  }
 },
 {
  "id": "2607.23006",
  "date": "2026-07-25",
  "title": "VecTree-RAG: An Agentic Retrieval-Augmented Generation Framework Combining Vector and Tree Retrieval for Efficiency and Accuracy",
  "scope_group": "other",
  "use": "literature-knowledge",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Chenghao-Wu/VecTree-RAG",
   "https://github.com/VikParuchuri/marker"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "DeepSeek-V4-Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "DeepSeek-V4-Flash",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "DeepSeek-V4-Pro",
    "version": null,
    "role": "judge"
   },
   {
    "name": "DeepSeek-V4-Pro",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "unspecified LLM",
    "version": null,
    "role": "other"
   },
   {
    "name": "SPECTER2",
    "version": "allenai/specter2_base",
    "role": "other"
   },
   {
    "name": "BGE cross-encoder",
    "version": "bge-reranker-v2-m3",
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "deepseek-v4-pro",
   "newest_release": "2026-04-24",
   "newest_class": "open-weights",
   "lag_days": 92,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 24
  }
 },
 {
  "id": "2607.24479",
  "date": "2026-07-27",
  "title": "Two-copy nondistillability of Werner states: sharp partial-trace inequalities and finite-copy extensions",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "other",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "ChatGPT 5.6 Sol (GPT-5.6 Sol)",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-sol-gpt-5.6-sol",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 18,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 21
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2607.24687",
  "date": "2026-07-27",
  "title": "Sharp continuity of quantum conditional entropy",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "other",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "ChatGPT 5.6 Sol",
    "version": "5.6 Sol",
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-sol",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 18,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 7
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2607.24714",
  "date": "2026-07-27",
  "title": "Efficient LLM-Generated Shuttling Compilers for Complex Trapped-Ion Architectures",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "application",
  "harness": "off-the-shelf: Claude Code",
  "harness_class": "off-the-shelf",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/BensonZhou1991/circuittransform/tree/master/inputs/QASM%20example",
   "https://github.com/teaguetomesh/quantum_circuit_generator"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Opus 4.7",
    "version": "4.7",
    "role": "system"
   },
   {
    "name": "Claude Fable 5",
    "version": "5",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-fable-5",
   "newest_release": "2026-06-09",
   "newest_class": "frontier-api",
   "lag_days": 48,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 86
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.25145",
  "date": "2026-07-27",
  "title": "Agentic AI for Scientific Reasoning in Autonomous Quantum Sensing Experiments",
  "scope_group": "condensed-matter",
  "use": "experimental-control",
  "kind": "application",
  "harness": "off-the-shelf: OpenClaw (release); Codex CLI (Table S2 benchmark logs)",
  "harness_class": "off-the-shelf",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/takuyaisogawa/nv-autonomous-experiments"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.6 Sol",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-sol",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 18,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 2,
   "bibliography_size": 24
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2607.25647",
  "date": "2026-07-28",
  "title": "KQFuzz: Knowledge-Guided Fuzzing for Quantum Libraries via Large Language Models",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 6,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://doi.org/10.5281/zenodo.19230459"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Gemini-3",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen2.5-Coder-3B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen2.5-Coder-7B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen2.5-Coder-14B",
    "version": null,
    "role": "system"
   },
   {
    "name": "CodeLlama-7B",
    "version": null,
    "role": "system"
   },
   {
    "name": "CodeLlama-13B",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "qwen2.5-coder-3b",
   "newest_release": "2024-11-12",
   "newest_class": "open-weights",
   "lag_days": 623,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 36
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 68
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.25865",
  "date": "2026-07-28",
  "title": "OmniQEC: discovering practical quantum error-correcting codes by an AI scientist",
  "scope_group": "adjacent-quantum",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/quantumlib/Stim"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Opus 4.8",
    "version": "4.8",
    "role": "system"
   },
   {
    "name": "DeepSeek-V4-Pro",
    "version": "V4-Pro",
    "role": "system"
   },
   {
    "name": "GPT-5.5",
    "version": "5.5",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.8",
   "newest_release": "2026-05-28",
   "newest_class": "frontier-api",
   "lag_days": 61,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 2,
   "bibliography_size": 91
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.26305",
  "date": "2026-07-28",
  "title": "Inclusions between p-bounded crystalline loci in dimension two",
  "scope_group": "other",
  "use": "reasoning-proof",
  "kind": "other",
  "harness": "none",
  "harness_class": "none",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/davidsavitt/KLSposet-artifacts",
   "https://github.com/davidsavitt/KLSposet"
  ],
  "ai_use_statement": true,
  "models": [],
  "model_lag": {
   "newest_model": "gpt-5.5-pro",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 96,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 8
  }
 },
 {
  "id": "2607.27262",
  "date": "2026-07-29",
  "title": "LLM-Guided Initialization for Accelerated Hybrid Quantum-Classical Medical Image Classification",
  "scope_group": "adjacent-quantum",
  "use": "design-optimization",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://nvidia.github.io/cuda-quantum/"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen2.5-3B-Instruct",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "qwen2.5-3b",
   "newest_release": "2024-09-19",
   "newest_class": "open-weights",
   "lag_days": 678,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 36
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 28
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2607.28200",
  "date": "2026-07-30",
  "title": "Vibe-FDTR: An agent-oriented framework for reproducible frequency-domain thermoreflectance data analysis",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "method",
  "harness": "off-the-shelf: OpenCode",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/yfwsunny/Vibe_FDTR",
   "https://github.com/lwh9346/Vibe_FDTR_benchmarks"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "DeepSeek-V4-Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "DeepSeek-V4-Pro",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "deepseek-v4-pro",
   "newest_release": "2026-04-24",
   "newest_class": "open-weights",
   "lag_days": 97,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 45
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "general",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2608.00149",
  "date": "2026-07-31",
  "title": "A Synthetically-accessible Universe of Chemically Recyclable Polymers",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Ramprasad-Group/polyVERSE/tree/main/ROP",
   "https://github.com/Ramprasad-Group/polyVERSE/tree/main/Virtual-Polymer/VFS/ROP/packaging_replacements",
   "https://github.com/Ramprasad-Group/canonicalize_psmiles"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "polyBART",
    "version": null,
    "role": "system"
   },
   {
    "name": "POLYT5",
    "version": null,
    "role": "system"
   },
   {
    "name": "SELFIES-TED",
    "version": null,
    "role": "other"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 57
  },
  "subject_codes": {
   "sol": {
    "subject": "molecular-soft",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "molecular-soft",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2608.01431",
  "date": "2026-08-02",
  "title": "PolymerGPT: Multi-property Optimization with a Decoder-Based GPT Model for Generative Polymer Design",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "none",
  "harness_class": "none",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://www.kaggle.com/code/awguhst/polymer-glass-transition-tg-autogluon/notebook"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "PolymerGPT",
    "version": null,
    "role": "system"
   },
   {
    "name": "PolyBERT",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "TransPolymer",
    "version": null,
    "role": "judge"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 15
  },
  "subject_codes": {
   "sol": {
    "subject": "molecular-soft",
    "question": "method-capability"
   },
   "opus": {
    "subject": "molecular-soft",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2608.02642",
  "date": "2026-07-31",
  "title": "MDArena: Evaluating Coding Agents on Realistic Molecular Dynamics Workflows",
  "scope_group": "other",
  "use": "simulation-workflow",
  "kind": "benchmark",
  "harness": "off-the-shelf: Harbor orchestrating Codex and OpenCode",
  "harness_class": "off-the-shelf",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/weitse-hsu/MDArena"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini-3.5-Flash",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Gemini-3.1-Pro-Preview",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "DeepSeek-V3.2-MAAS",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3-235B-A22B-Instruct-2507-MAAS",
    "version": "2507",
    "role": "baseline"
   },
   {
    "name": "Claude Opus 4.8",
    "version": null,
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.8",
   "newest_release": "2026-05-28",
   "newest_class": "frontier-api",
   "lag_days": 64,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 45
  }
 },
 {
  "id": "2608.03824",
  "date": "2026-08-04",
  "title": "Impossibility of Perfectly Complete Many-Round Key Agreement in the QROM",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "other",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT-5.6 Sol Ultra",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-sol-ultra",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 26,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 11
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2608.04947",
  "date": "2026-08-05",
  "title": "Sharp Continuity of Petz and Sandwiched Rényi Conditional Entropies",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "other",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "ChatGPT 5.6 Sol",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-sol",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 27,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 41
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2608.05266",
  "date": "2026-08-05",
  "title": "Agentic self-driving microscopy benchmarks support qualification but do not necessarily generalize to unseen tasks",
  "scope_group": "materials-science",
  "use": "experimental-control",
  "kind": "benchmark",
  "harness": "custom: LangChain/LangGraph ReAct, MCP",
  "harness_class": "custom",
  "base_models": 5,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/natertott/agentic_microscopy_benchmarks_XRM"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Sonnet 3.7",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Sonnet 4",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Opus 4.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-4.1",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.1",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.5",
   "newest_release": "2025-11-24",
   "newest_class": "frontier-api",
   "lag_days": 254,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 15
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 37
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2608.05387",
  "date": "2026-08-05",
  "title": "Fine-Tuning Small Language Models for Reliable VASP INCAR Generation",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 10,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/zxy-boop/VASPINCAR-Gen",
   "https://github.com/liuzhiyangnku/incarbench"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen3-0.6B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-1.7B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-4B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-8B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Llama-3.2-1B-Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "Llama-3.2-3B-Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "Llama-3.1-8B-Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemma-3-1B-IT",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemma-3-4B-IT",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemma-3-12B-IT",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "DeepSeek V4 Pro",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3.6-Plus",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "deepseek-v4-pro",
   "newest_release": "2026-04-24",
   "newest_class": "open-weights",
   "lag_days": 103,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 40
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2608.06694",
  "date": "2026-08-07",
  "title": "A Multi-Agent Framework for Automated Coarse-Grained Molecular Dynamics of Polymers",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 5,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "5.4-nano",
    "version": null,
    "role": "system"
   },
   {
    "name": "gpt-5.4-mini",
    "version": "gpt-5.4-mini",
    "role": "system"
   },
   {
    "name": "5.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Haiku 4.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Sonnet 5",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-5",
   "newest_release": "2026-06-30",
   "newest_class": "frontier-api",
   "lag_days": 38,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 46
  },
  "subject_codes": {
   "sol": {
    "subject": "molecular-soft",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "molecular-soft",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2608.07439",
  "date": "2026-08-07",
  "title": "An Exploratory Evaluation of LLM-Assisted Rewriting of Moderate-Complexity Financial Sentences for DisCoCat-Based Sentiment Analysis",
  "scope_group": "other",
  "use": "unclear",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/bllin001/qnlp-discocat-llms-finance-rewriting"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT-4.1-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen2.5:7B",
    "version": "2.5:7B",
    "role": "system"
   },
   {
    "name": "ChatGPT",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "MeaningBERT",
    "version": null,
    "role": "judge"
   },
   {
    "name": "FinBERT",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Google’s Gemini AI",
    "version": null,
    "role": "writing"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-4.1-mini",
   "newest_release": "2025-04-14",
   "newest_class": "frontier-api",
   "lag_days": 480,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 30
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 39
  }
 },
 {
  "id": "2608.07637",
  "date": "2026-08-07",
  "title": "Agent-MD: Selective LLM Intervention with Event-Driven Escalation for Stateful GCMC--MD Campaigns",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "application",
  "harness": "off-the-shelf: Codex CLI",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/yj-wang14/Agent-MD"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 106,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 31
  },
  "subject_codes": {
   "sol": {
    "subject": "functional-solids",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "functional-solids",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2608.08752",
  "date": "2026-08-09",
  "title": "Overlap distribution of the critical Sherrington-Kirkpatrick model",
  "scope_group": "condensed-matter",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT-5.6 Pro",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-pro",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 31,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 36
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2608.10483",
  "date": "2026-08-11",
  "title": "Predicting Space Groups of Double Perovskites by LLM with Dynamic Few-Shot Learning",
  "scope_group": "materials-science",
  "use": "property-prediction",
  "kind": "method",
  "harness": "off-the-shelf: Microsoft AutoGen",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.4-mini",
    "version": "gpt-5.4-mini-2026-03-17",
    "role": "system"
   },
   {
    "name": "LLM (unspecified)",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4-mini",
   "newest_release": "2026-03-17",
   "newest_class": "frontier-api",
   "lag_days": 147,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 8
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "functional-solids",
    "question": "method-capability"
   },
   "opus": {
    "subject": "functional-solids",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2608.10882",
  "date": "2026-08-11",
  "title": "From Pattern Detection to Composition Analysis in Quantum Software",
  "scope_group": "adjacent-quantum",
  "use": "literature-knowledge",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 4,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated; repository search July 21, 2026",
  "code_urls": [
   "https://github.com/saeg/qpa-v2"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "meta-llama/Llama-3.3-70B-Instruct-Turbo",
    "version": "meta-llama/Llama-3.3-70B-Instruct-Turbo",
    "role": "system"
   },
   {
    "name": "Qwen/Qwen2.5-7B-Instruct-Turbo",
    "version": "Qwen/Qwen2.5-7B-Instruct-Turbo",
    "role": "system"
   },
   {
    "name": "deepseek-ai/DeepSeek-V4-Pro",
    "version": "deepseek-ai/DeepSeek-V4-Pro",
    "role": "system"
   },
   {
    "name": "fourth model (unnamed)",
    "version": null,
    "role": "system"
   },
   {
    "name": "all-mpnet-base-v2",
    "version": "all-mpnet-base-v2",
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "deepseek-v4-pro",
   "newest_release": "2026-04-24",
   "newest_class": "open-weights",
   "lag_days": 109,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 40
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2608.11224",
  "date": "2026-07-25",
  "title": "Harnessing agent memory to build lifelong AI partners for materials scientists",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 5,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.2",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4-nano",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3.5-397B-A17B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini-3.1-Pro",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3-Max",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Claude-Sonnet-4.6",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "qwen3-max",
    "version": null,
    "role": "other"
   },
   {
    "name": "qwen3-embedding-8b",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4-nano",
   "newest_release": "2026-03-17",
   "newest_class": "frontier-api",
   "lag_days": 130,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 8
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 55
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2608.11283",
  "date": "2026-08-11",
  "title": "Chemically Meaningful Textualization Enables Explainable Validation of Metal-Organic Frameworks by Large Language Models",
  "scope_group": "materials-science",
  "use": "property-prediction",
  "kind": "method",
  "harness": "none",
  "harness_class": "none",
  "base_models": 4,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/sxm13/MOF2Text",
   "https://github.com/sxm13/MOFClassifier2"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Llama3-8B",
    "version": null,
    "role": "system"
   },
   {
    "name": "DeepSeek-7B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemma3-4B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-4B",
    "version": null,
    "role": "system"
   },
   {
    "name": "LLM-prop",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "qwen3-4b",
   "newest_release": "2025-04-29",
   "newest_class": "open-weights",
   "lag_days": 469,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 29
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "porous-frameworks",
    "question": "method-capability"
   },
   "opus": {
    "subject": "porous-frameworks",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2608.15900",
  "date": "2026-08-16",
  "title": "Crystal-structure design by agentic AI in a language of motifs",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/minhquyet2308/MatEvolve"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Fable 5",
    "version": "claude-fable-5",
    "role": "system"
   },
   {
    "name": "Claude Fable 5",
    "version": "claude-fable-5",
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "claude-fable-5",
   "newest_release": "2026-06-09",
   "newest_class": "frontier-api",
   "lag_days": 68,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 166
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-materials",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-materials",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2608.15928",
  "date": "2026-08-16",
  "title": "Synthesizing like a chemist: an iterative, feedback-driven loop for materials discovery",
  "scope_group": "materials-science",
  "use": "literature-knowledge",
  "kind": "application",
  "harness": "off-the-shelf: PaperQA via Ollama",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/PV-Lab/SpinSynth",
   "https://github.com/PV-Lab/MOBO-Kit"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Qwen2.5-14B",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "qwen2.5-14b",
   "newest_release": "2024-09-19",
   "newest_class": "open-weights",
   "lag_days": 696,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 37
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "semiconductor-optical",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "semiconductor-optical",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2608.16048",
  "date": "2026-08-17",
  "title": "zenDot: An LLM-integrated quantum TCAD platform for semiconductor quantum-device design and optimization automation",
  "scope_group": "condensed-matter",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "one local language model",
    "version": null,
    "role": "system"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2608.17629",
  "date": "2026-08-18",
  "title": "Unclonable encryption from BB84 states: a simultaneous Goldreich-Levin reduction",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT-5.6 Ultra",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-ultra",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 40,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 2
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 27
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2608.19116",
  "date": "2026-08-19",
  "title": "JANUS: A Multi-modal Foundation Neural Sampler for Disordered Materials",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "LLM",
    "version": null,
    "role": "system"
   }
  ],
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 79
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2608.19790",
  "date": "2026-08-20",
  "title": "LLMs as Acquisition Policies for Finite-Pool Materials Optimization: A Controlled Study",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 5,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Gemma 4 31B",
    "version": null,
    "role": "system"
   },
   {
    "name": "DeepSeek-V4-Flash (284B-A13B)",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3.5-27B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3.5-35B-A3B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3.5-397B-A17B",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "deepseek-v4-flash",
   "newest_release": "2026-04-24",
   "newest_class": "open-weights",
   "lag_days": 118,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 7
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 35
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2608.20434",
  "date": "2026-08-20",
  "title": "An LLM agent for end-to-end computational materials discovery",
  "scope_group": "materials-science",
  "use": "research-orchestration",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Opus 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "Claude Opus 4.6",
    "version": "4.6",
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.6",
   "newest_release": "2026-02-05",
   "newest_class": "frontier-api",
   "lag_days": 196,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 13
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "porous-frameworks",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "porous-frameworks",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2608.23631",
  "date": "2026-08-23",
  "title": "TRACE: Transition-Aware Residual Control for Multi-Objective Materials Discovery",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen2.5-14B-Instruct",
    "version": "Qwen2.5-14B-Instruct",
    "role": "system"
   },
   {
    "name": "Mistral-Small-3.2-24B-Instruct-2506",
    "version": "Mistral-Small-3.2-24B-Instruct-2506",
    "role": "system"
   },
   {
    "name": "Qwen2.5-14B-Instruct",
    "version": "Qwen2.5-14B-Instruct",
    "role": "baseline"
   },
   {
    "name": "Mistral-Small-3.2-24B-Instruct-2506",
    "version": "Mistral-Small-3.2-24B-Instruct-2506",
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "mistral-small-3.2-24b-2506",
   "newest_release": "2025-06-20",
   "newest_class": "open-weights",
   "lag_days": 429,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 26
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 39
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2608.25189",
  "date": "2026-08-25",
  "title": "What Should a Large Language Model See? Physical Invariants as a Data Representation for PDE Discovery",
  "scope_group": "other",
  "use": "reasoning-proof",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "QwQ-32B",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "qwq-32b",
   "newest_release": "2025-03-06",
   "newest_class": "open-weights",
   "lag_days": 537,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 33
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 9
  }
 },
 {
  "id": "2608.25392",
  "date": "2026-08-26",
  "title": "Interpretable physics-informed retrieval-augmented generation language model for end-to-end inorganic crystal synthesis planning",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Jweijian/PIRAG-LM",
   "https://groupcookie.com"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "DeepSeek-V4",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "LLM",
    "version": null,
    "role": "system"
   },
   {
    "name": "LLM-only",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Q&A-LLM",
    "version": null,
    "role": "system"
   },
   {
    "name": "large language models",
    "version": null,
    "role": "judge"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 40
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2608.27363",
  "date": "2026-08-27",
  "title": "On supporting affine functionals for Entanglement of Formation",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Fable 5",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-fable-5",
   "newest_release": "2026-06-09",
   "newest_class": "frontier-api",
   "lag_days": 79,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 21
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2608.27572",
  "date": "2026-08-27",
  "title": "Scale-free correlations in extremely persistent active matter",
  "scope_group": "condensed-matter",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Opus 4.8",
    "version": "4.8",
    "role": "system"
   },
   {
    "name": "Opus 5",
    "version": "5",
    "role": "system"
   },
   {
    "name": "Fable 5",
    "version": "5",
    "role": "system"
   },
   {
    "name": "Gemini AI",
    "version": null,
    "role": "writing"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-5",
   "newest_release": "2026-07-24",
   "newest_class": "frontier-api",
   "lag_days": 34,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 21
  },
  "subject_codes": {
   "sol": {
    "subject": "molecular-soft",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "molecular-soft",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2608.27587",
  "date": "2026-08-27",
  "title": "Compiling Chemical Knowledge into Executable Descriptors for Materials Prediction",
  "scope_group": "materials-science",
  "use": "property-prediction",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated; temporal data: October 2022–April 2026",
  "code_urls": [
   "https://github.com/snu-micc/CRISP"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "gpt-4.1-mini",
    "version": "gpt-4.1-mini-2025-04-14",
    "role": "system"
   },
   {
    "name": "GPT-5",
    "version": null,
    "role": "system"
   },
   {
    "name": "gpt-4.1",
    "version": "gpt-4.1-2025-04-14",
    "role": "system"
   },
   {
    "name": "text-embedding-3-large",
    "version": null,
    "role": "other"
   },
   {
    "name": "GPT-4o",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Llama-3-8B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "CSLLM",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "StructLLM",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "SynCry",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5",
   "newest_release": "2025-08-07",
   "newest_class": "frontier-api",
   "lag_days": 385,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 22
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2608.28100",
  "date": "2026-08-28",
  "title": "uMOF: A Universal Database, Benchmark, and Machine Learning Interatomic Potentials for Metal-Organic Frameworks",
  "scope_group": "materials-science",
  "use": "literature-knowledge",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://doi.org/10.6084/m9.figshare.33311829",
   "https://github.com/ddmms/ml-peg"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Haiku 4.5",
    "version": "4.5",
    "role": "system"
   },
   {
    "name": "Claude Opus 4.7",
    "version": "4.7",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.7",
   "newest_release": "2026-04-16",
   "newest_class": "frontier-api",
   "lag_days": 134,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 8
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 93
  },
  "subject_codes": {
   "sol": {
    "subject": "porous-frameworks",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "porous-frameworks",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2608.29309",
  "date": "2026-08-29",
  "title": "Evaluating LLM-based AI agents integrated with materials synthesis tools: the case of atomic layer deposition",
  "scope_group": "materials-science",
  "use": "agent-evaluation",
  "kind": "review-perspective",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-3.5",
    "version": null,
    "role": "other"
   },
   {
    "name": "GPT-4.0",
    "version": null,
    "role": "other"
   },
   {
    "name": "GPT-4o",
    "version": null,
    "role": "other"
   },
   {
    "name": "o1",
    "version": null,
    "role": "other"
   },
   {
    "name": "o3",
    "version": null,
    "role": "other"
   },
   {
    "name": "GPT5",
    "version": null,
    "role": "other"
   },
   {
    "name": "Claude Sonnet 4",
    "version": null,
    "role": "other"
   },
   {
    "name": "Claude Opus 4",
    "version": null,
    "role": "other"
   },
   {
    "name": "Claude Sonnet 4.5",
    "version": null,
    "role": "other"
   },
   {
    "name": "Gemini 2.5 Flash",
    "version": null,
    "role": "other"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 38
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "ai-workflow-perspective"
   },
   "opus": {
    "subject": "general",
    "question": "ai-workflow-perspective"
   }
  }
 },
 {
  "id": "2608.29665",
  "date": "2026-08-30",
  "title": "Evaluating a 4B open-weights local LLM for agentic DFT workflows: a literature reproducibility audit",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "application",
  "harness": "custom (Ollama 0.32.11 / llama.cpp)",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "Not stated; retrieval: 16 August 2026",
  "code_urls": [
   "https://github.com/shambhubhandari/Agentic-Workflow"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "qwen3:4b",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "qwen3-4b",
   "newest_release": "2025-04-29",
   "newest_class": "open-weights",
   "lag_days": 488,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 30
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 36
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2609.00070",
  "date": "2026-08-30",
  "title": "AutoXRD: Autonomous LLM Agents and Comprehensive Evaluation for Powder Diffraction Analysis",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 10,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://stephen-smj.github.io/AutoXRD/"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.6 Sol",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.6 Sol",
    "version": null,
    "role": "judge"
   },
   {
    "name": "GPT-5.6 Terra",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.6 Luna",
    "version": null,
    "role": "system"
   },
   {
    "name": "MiMo-V2.5-Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "MiMo-V2.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "DeepSeek-V4-Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "DeepSeek-V4-Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "MiniMax-M2.7",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Haiku 4.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini 3.1 Pro Low",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-terra",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 52,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 2
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 1,
   "bibliography_size": 50
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2609.00099",
  "date": "2026-08-31",
  "title": "Generative artificial intelligence for reliable mechanistic reasoning for corrosion",
  "scope_group": "materials-science",
  "use": "scientific-qa",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Llama-3.1-8B-Instruct",
    "version": "3.1",
    "role": "system"
   },
   {
    "name": "Qwen-2.5-7B-Instruct",
    "version": "2.5",
    "role": "system"
   },
   {
    "name": "Mistral-7B-Instruct-v0.3",
    "version": "v0.3",
    "role": "system"
   },
   {
    "name": "GPT-4o-mini",
    "version": null,
    "role": "judge"
   },
   {
    "name": "unnamed Q-A LLM",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "unnamed proposition-extraction LLM",
    "version": null,
    "role": "other"
   },
   {
    "name": "DeBERTa-v3-large",
    "version": "v3-large",
    "role": "judge"
   },
   {
    "name": "SPECTER2",
    "version": null,
    "role": "other"
   },
   {
    "name": "BGE-M3",
    "version": null,
    "role": "other"
   },
   {
    "name": "unspecified BERTScore encoder",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "qwen2.5-7b",
   "newest_release": "2024-09-19",
   "newest_class": "open-weights",
   "lag_days": 711,
   "frontier_at_paper": "Grok 4.6 (2026-08-12)",
   "flagship_releases_since": 37
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "alloys",
    "question": "method-capability"
   },
   "opus": {
    "subject": "alloys",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2609.00663",
  "date": "2026-09-01",
  "title": "Separating perception from reasoning in vision-language models: a model-free render ceiling for crystal structures",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 17,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/KurbanIntelligenceLab/render-ceiling"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "google/gemini-3.6-flash",
    "version": "google/gemini-3.6-flash",
    "role": "system"
   },
   {
    "name": "x-ai/grok-4.5",
    "version": "x-ai/grok-4.5",
    "role": "system"
   },
   {
    "name": "anthropic/claude-opus-4.8",
    "version": "anthropic/claude-opus-4.8",
    "role": "system"
   },
   {
    "name": "meta-llama/llama-4-maverick",
    "version": "meta-llama/llama-4-maverick",
    "role": "system"
   },
   {
    "name": "z-ai/glm-4.6v",
    "version": "z-ai/glm-4.6v",
    "role": "system"
   },
   {
    "name": "qwen/qwen3-vl-8b-instruct",
    "version": "qwen/qwen3-vl-8b-instruct",
    "role": "system"
   },
   {
    "name": "openai/gpt-4.1-mini",
    "version": "openai/gpt-4.1-mini",
    "role": "system"
   },
   {
    "name": "qwen/qwen3-vl-235b-a22b-instruct",
    "version": "qwen/qwen3-vl-235b-a22b-instruct",
    "role": "system"
   },
   {
    "name": "bytedance-seed/seed-1.6",
    "version": "bytedance-seed/seed-1.6",
    "role": "system"
   },
   {
    "name": "mistralai/mistral-medium-3.1",
    "version": "mistralai/mistral-medium-3.1",
    "role": "system"
   },
   {
    "name": "qwen/qwen3-vl-32b-instruct",
    "version": "qwen/qwen3-vl-32b-instruct",
    "role": "system"
   },
   {
    "name": "meta-llama/llama-4-scout",
    "version": "meta-llama/llama-4-scout",
    "role": "system"
   },
   {
    "name": "amazon/nova-pro-v1",
    "version": "amazon/nova-pro-v1",
    "role": "system"
   },
   {
    "name": "mistralai/mistral-small-2603",
    "version": "mistralai/mistral-small-2603",
    "role": "system"
   },
   {
    "name": "qwen/qwen2.5-vl-72b-instruct",
    "version": "qwen/qwen2.5-vl-72b-instruct",
    "role": "system"
   },
   {
    "name": "Gemini 2.5 Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Opus 5",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-5",
   "newest_release": "2026-07-24",
   "newest_class": "frontier-api",
   "lag_days": 39,
   "frontier_at_paper": "Claude Fable 5.1 (2026-09-01)",
   "flagship_releases_since": 2
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 45
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2609.00795",
  "date": "2026-09-01",
  "title": "Agentic programs: an emerging form of scientific software in computational materials science",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "review-perspective",
  "harness": "off-the-shelf: Claude Code",
  "harness_class": "off-the-shelf",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/anthropics/claude-code",
   "https://github.com/openai/codex"
  ],
  "ai_use_statement": false,
  "models": [],
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 34
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "ai-workflow-perspective"
   },
   "opus": {
    "subject": "general",
    "question": "ai-workflow-perspective"
   }
  }
 },
 {
  "id": "2609.04674",
  "date": "2026-09-04",
  "title": "QMClaw: A Scalable General-purpose Framework for Quantum Measurement and Control",
  "scope_group": "adjacent-quantum",
  "use": "experimental-control",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "LLM",
    "version": null,
    "role": "system"
   },
   {
    "name": "LLM-based sequence predictor",
    "version": null,
    "role": "baseline"
   }
  ],
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 54
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2609.05039",
  "date": "2026-09-04",
  "title": "Qlippy: A Retrieval-Augmented GenAI Assistant for Reproducible Quantum Workflows and Experiment Tracking",
  "scope_group": "adjacent-quantum",
  "use": "scientific-qa",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 4,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "Judge API accessed April 2026; generator dates not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-4o-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemma4:e4b",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3.5:9b",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4-mini",
    "version": null,
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "gemma-4-e4b",
   "newest_release": "2026-04-02",
   "newest_class": "open-weights",
   "lag_days": 155,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 11
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 10
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2609.05121",
  "date": "2026-09-04",
  "title": "On the growth of operator entanglement in brickwork circuits with Yang--Baxter gates",
  "scope_group": "condensed-matter",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "ChatGPT 5.6 Sol",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-sol",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 57,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 50
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2609.05157",
  "date": "2026-09-04",
  "title": "AxQM: A Textbook-Scale Benchmark for Formal Proof Synthesis in a Library of Finite-Dimensional Quantum Mechanics",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Axiomatic-AI/AxQM"
  ],
  "ai_use_statement": false,
  "models": [],
  "citations": {
   "cites_in_set": 4,
   "cited_by_in_set": 0,
   "bibliography_size": 77
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2609.06038",
  "date": "2026-09-05",
  "title": "The $I_{3322}$ Bell inequality requires infinite dimensions",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT 5.5 Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT 5.6 Sol",
    "version": null,
    "role": "writing"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5-pro",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 135,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 9
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 16
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2609.07655",
  "date": "2026-09-07",
  "title": "Online Surrogate Repair: Decoupling High-Fidelity Feedback from Search Length in Closed-Loop Discovery",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "DeepSeek-V4-Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.6-Luna",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-luna",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 60,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 51
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2609.10520",
  "date": "2026-09-09",
  "title": "Private communication via zero-private-capacity quantum channels",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "other",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/QudeLeap/lean-private-capacity"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "large language models",
    "version": null,
    "role": "system"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 23
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2609.12181",
  "date": "2026-09-10",
  "title": "Physics as the label for measuring and correcting materials reasoning in multimodal models",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 9,
  "dated_system_snapshot": false,
  "eval_dates_stated": true,
  "evaluation_dates": "2026-07-04–11; corrected reruns 2026-07-16/17; refresh 2026-07-19/20; flagship 2026-08-06",
  "code_urls": [
   "https://github.com/KurbanIntelligenceLab/matpcr"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Qwen2.5-VL-7B-Instruct",
    "version": "Qwen/Qwen2.5-VL-7B-Instruct",
    "role": "system"
   },
   {
    "name": "Gemini 2.5 Flash Lite",
    "version": "google/gemini-2.5-flash-lite",
    "role": "system"
   },
   {
    "name": "GPT-4o",
    "version": "openai/gpt-4o",
    "role": "system"
   },
   {
    "name": "Claude Sonnet 4.5",
    "version": "anthropic/claude-sonnet-4.5",
    "role": "system"
   },
   {
    "name": "Claude Opus 4.6",
    "version": "anthropic/claude-opus-4.6",
    "role": "system"
   },
   {
    "name": "GPT-5.6-Terra",
    "version": "openai/gpt-5.6-terra",
    "role": "system"
   },
   {
    "name": "Qwen3-VL-8B-Instruct",
    "version": "qwen/qwen3-vl-8b-instruct",
    "role": "system"
   },
   {
    "name": "Qwen3-VL-8B-Thinking",
    "version": "qwen/qwen3-vl-8b-thinking",
    "role": "system"
   },
   {
    "name": "GPT-5.6-Sol",
    "version": "openai/gpt-5.6-sol",
    "role": "system"
   },
   {
    "name": "MatterChat",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Mistral-7B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen2.5-0.5B-Instruct",
    "version": null,
    "role": "other"
   },
   {
    "name": "Gemini 2.5 Flash Lite",
    "version": "google/gemini-2.5-flash-lite",
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-terra",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 63,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 55
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2609.12184",
  "date": "2026-09-10",
  "title": "Agentic TCAD Calibration Workflow for Oxide Semiconductor Transistors",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "application",
  "harness": "off-the-shelf: OpenAI Codex and NVIDIA NeMo Agent Toolkit (reference specifies 1.8.0)",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/NVIDIA/NeMo-Agent-Toolkit/releases/tag/v1.8.0"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "OpenAI GPT-5.5",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 140,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 9
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 15
  },
  "subject_codes": {
   "sol": {
    "subject": "semiconductor-optical",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "semiconductor-optical",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2609.12796",
  "date": "2026-09-11",
  "title": "Agent-authored deposition recipes for X-ray multilayer mirrors: schema-bound LLM control of a magnetron sputtering system with reflectivity-verified outcomes",
  "scope_group": "materials-science",
  "use": "experimental-control",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Opus 5",
    "version": "5",
    "role": "system"
   },
   {
    "name": "Claude Fable 5.1",
    "version": "5.1",
    "role": "system"
   },
   {
    "name": "supervising LLM session (unnamed)",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "claude-fable-5.1",
   "newest_release": "2026-09-01",
   "newest_class": "frontier-api",
   "lag_days": 10,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 3,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "semiconductor-optical",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "semiconductor-optical",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2609.13121",
  "date": "2026-09-11",
  "title": "Optimal relaxation for Witten Lindbladians",
  "scope_group": "other",
  "use": "reasoning-proof",
  "kind": "other",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Chat GPT 6",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "chat-gpt-6",
   "newest_release": "2026-09-04",
   "newest_class": "frontier-api",
   "lag_days": 7,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 0
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 25
  }
 },
 {
  "id": "2609.13357",
  "date": "2026-09-11",
  "title": "Can Autonomous LLM Agents Execute Multireference Quantum Chemistry Calculations?",
  "scope_group": "other",
  "use": "simulation-workflow",
  "kind": "application",
  "harness": "off-the-shelf: Claude Code",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Opus 5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Microsoft Copilot (institutional instance)",
    "version": null,
    "role": "writing"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-5",
   "newest_release": "2026-07-24",
   "newest_class": "frontier-api",
   "lag_days": 49,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 134
  }
 },
 {
  "id": "2609.14080",
  "date": "2026-09-12",
  "title": "Adapting Open-Weight MLLMs to Generate Point Prompts for Electron Microscopy Segmentation",
  "scope_group": "other",
  "use": "data-interpretation",
  "kind": "method",
  "harness": "none",
  "harness_class": "none",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen3-VL",
    "version": "Qwen/Qwen3-VL-8B-Instruct",
    "role": "system"
   },
   {
    "name": "InternVL3.5",
    "version": "OpenGVLab/InternVL3_5-8B-HF",
    "role": "system"
   },
   {
    "name": "Molmo2",
    "version": "allenai/Molmo2-8B",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "molmo2",
   "newest_release": "2025-12-14",
   "newest_class": "open-weights",
   "lag_days": 272,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 16
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 45
  }
 },
 {
  "id": "2609.14840",
  "date": "2026-09-13",
  "title": "El Agente Potente: High-Throughput Agentic Atomistic Simulations",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom; baseline off-the-shelf: Claude Code with AtomisticSkills",
  "harness_class": "off-the-shelf",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": true,
  "evaluation_dates": "Not explicitly stated; C.5 log filename contains 20260609_181646",
  "code_urls": [
   "https://github.com/kenko911/User_Cases_El_Agente_Potente",
   "https://gitlab.com/ase/ase",
   "https://github.com/materialsproject/pymatgen",
   "https://github.com/ACEsuit/mace",
   "https://github.com/microsoft/mattersim",
   "https://github.com/orbital-materials/orb-models",
   "https://github.com/materialyzeai/matcalc",
   "https://github.com/pfnet-research/torch-dftd",
   "https://github.com/lab-cosmo/upet",
   "https://github.com/phonopy/phonopy",
   "https://github.com/phonopy/phono3py",
   "https://github.com/learningmatter-mit/AtomisticSkills/blob/ca695e1b89ea48613291c88bd8c73e71af2025c7/src/utils/mlips/mace/mace_wrapper.py#L18"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "ChatGPT 5.5",
    "version": "5.5",
    "role": "system"
   },
   {
    "name": "ChatGPT 4.1",
    "version": "4.1",
    "role": "system"
   },
   {
    "name": "sonnet-4.6",
    "version": "sonnet-4.6",
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 143,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 9
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 1,
   "bibliography_size": 72
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "general",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2609.14928",
  "date": "2026-09-14",
  "title": "Quantitative control and recording of materials-synthesis processes using an automated experimentation platform",
  "scope_group": "materials-science",
  "use": "experimental-control",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/YusukeHashimotoLab/saigen"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Gemini 3.5 Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "whisper-large-v3-turbo",
    "version": "whisper-large-v3-turbo",
    "role": "other"
   },
   {
    "name": "YOLOv8",
    "version": "v8",
    "role": "other"
   },
   {
    "name": "Claude",
    "version": null,
    "role": "writing"
   }
  ],
  "model_lag": {
   "newest_model": "gemini-3.5-flash",
   "newest_release": "2026-05-19",
   "newest_class": "frontier-api",
   "lag_days": 118,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 8
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 1,
   "bibliography_size": 39
  },
  "subject_codes": {
   "sol": {
    "subject": "porous-frameworks",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "general",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2609.16466",
  "date": "2026-09-15",
  "title": "Scaling LLM Agents for Materials Design through Hierarchical Collective Reasoning",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "Exact dates not stated; architecture and reported agent trajectories completed before 8 September 2026.",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.6 Terra",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.6 Sol",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-6 Astra",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-6-astra",
   "newest_release": "2026-09-04",
   "newest_class": "frontier-api",
   "lag_days": 11,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 0
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2609.17123",
  "date": "2026-09-15",
  "title": "AI for Science with GPT-6 Astra: Thermal Design and Electrothermal Analysis of 2D CFET",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-6",
    "version": null,
    "role": "other"
   },
   {
    "name": "GPT-6 Astra",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Fable 5.1",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Opus 5",
    "version": null,
    "role": "other"
   },
   {
    "name": "GPT-5.6 Sol",
    "version": null,
    "role": "other"
   },
   {
    "name": "GPT-5.6 Terra",
    "version": null,
    "role": "other"
   },
   {
    "name": "GPT-5 / Codex",
    "version": null,
    "role": "other"
   },
   {
    "name": "Claude Haiku 4.5",
    "version": null,
    "role": "other"
   },
   {
    "name": "Claude Opus 4.8",
    "version": null,
    "role": "other"
   },
   {
    "name": "Claude Sonnet 5",
    "version": null,
    "role": "other"
   },
   {
    "name": "GPT-5.6 Luna",
    "version": null,
    "role": "other"
   },
   {
    "name": "Kimi",
    "version": null,
    "role": "other"
   },
   {
    "name": "Perplexity [Grok 4.6]",
    "version": null,
    "role": "other"
   },
   {
    "name": "Grok-labelled table",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-6-astra",
   "newest_release": "2026-09-04",
   "newest_class": "frontier-api",
   "lag_days": 11,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 0
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 17
  },
  "subject_codes": {
   "sol": {
    "subject": "semiconductor-optical",
    "question": "method-capability"
   },
   "opus": {
    "subject": "semiconductor-optical",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2609.17291",
  "date": "2026-09-15",
  "title": "Extracting ontology-compliant knowledge from scientific text describing irradiated materials using large language models",
  "scope_group": "materials-science",
  "use": "literature-knowledge",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 6,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/jmendi/eolas_irradiated_materials/"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "gpt-oss-120b",
    "version": "gpt-oss-120b",
    "role": "system"
   },
   {
    "name": "gpt-oss-20b",
    "version": "gpt-oss-20b",
    "role": "system"
   },
   {
    "name": "granite-3.3-8b-instruct",
    "version": "granite-3.3-8b-instruct",
    "role": "system"
   },
   {
    "name": "llama-3-3-70b-instruct",
    "version": "Llama-3.3-70B-Instruct",
    "role": "system"
   },
   {
    "name": "llama-4-maverick-17b-128e-instruct-fp8",
    "version": "Llama-4-Maverick-17B-128E-Instruct-FP8",
    "role": "system"
   },
   {
    "name": "mistral-large",
    "version": "Mistral-Large-Instruct-2407",
    "role": "system"
   },
   {
    "name": "unspecified LLM",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "unspecified LLM",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-oss-20b",
   "newest_release": "2025-08-05",
   "newest_class": "open-weights",
   "lag_days": 406,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 25
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 64
  },
  "subject_codes": {
   "sol": {
    "subject": "structural-composites",
    "question": "method-capability"
   },
   "opus": {
    "subject": "structural-composites",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2609.18598",
  "date": "2026-09-16",
  "title": "Hypothesis-Driven Autonomous Materials Synthesis with Multimodal LLM Agents",
  "scope_group": "materials-science",
  "use": "research-orchestration",
  "kind": "application",
  "harness": "off-the-shelf: OpenAI Agents SDK",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT-5.5",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 146,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 9
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 41
  },
  "subject_codes": {
   "sol": {
    "subject": "battery-materials",
    "question": "method-capability"
   },
   "opus": {
    "subject": "battery-materials",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2609.19550",
  "date": "2026-09-17",
  "title": "Portfolio-Based Constrained Multi-Objective Bayesian Optimization for Materials Design",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/sushant-k-sinha/PCMOBO"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "gpt-4o",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-4o",
   "newest_release": "2024-05-13",
   "newest_class": "frontier-api",
   "lag_days": 857,
   "frontier_at_paper": "GPT-6 Astra (2026-09-04)",
   "flagship_releases_since": 39
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 30
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2609.19837",
  "date": "2026-09-17",
  "title": "Machine Learning for High-Entropy Catalysts: Methods and Applications",
  "scope_group": "materials-science",
  "use": "literature-knowledge",
  "kind": "review-perspective",
  "harness": "none",
  "harness_class": "none",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-4",
    "version": null,
    "role": "other"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 145
  },
  "subject_codes": {
   "sol": {
    "subject": "catalysts",
    "question": "ai-workflow-perspective"
   },
   "opus": {
    "subject": "catalysts",
    "question": "ai-workflow-perspective"
   }
  }
 },
 {
  "id": "2609.20258",
  "date": "2026-07-28",
  "title": "A Protocol-Guided LLM Agent for Quantum Program Synthesis and Execution",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "application",
  "harness": "off-the-shelf: OpenAI Agents SDK; custom Qiskit evaluation",
  "harness_class": "off-the-shelf",
  "base_models": 4,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Gemma-4-31B-IT",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3.6-35B-A3B",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-OSS-120B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Nemotron-3-Ultra-550B-A55B",
    "version": null,
    "role": "system"
   },
   {
    "name": "ChatGPT",
    "version": null,
    "role": "writing"
   }
  ],
  "model_lag": {
   "newest_model": "nemotron-3-ultra-550b-a55b",
   "newest_release": "2026-06-04",
   "newest_class": "open-weights",
   "lag_days": 54,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 4
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 13
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2609.20319",
  "date": "2026-07-31",
  "title": "QEncodeBench: Can Large Language Models Encode Classical Problems into Verified Quantum Oracles?",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "benchmark",
  "harness": "custom; off-the-shelf: Claude Code, Codex; vLLM serving",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/chexujun/QEncodeBench"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "DeepSeek-V4-Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen2.5-Coder-7B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen2.5-Coder-32B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Qwen3-32B",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "DeepSeek-R1-0528-Qwen3-8B",
    "version": "0528",
    "role": "baseline"
   },
   {
    "name": "Qwen3-235B-A22B-Thinking",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "gpt-oss-20b",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "gpt-oss-120b",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Claude Opus 4.8",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Claude Haiku 4.5",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-5.6 Terra",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-terra",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 22,
   "frontier_at_paper": "Claude Opus 5 (2026-07-24)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 42
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2609.22210",
  "date": "2026-09-01",
  "title": "SALSA: Semi-Autonomous Literature Summarization Assistant",
  "scope_group": "materials-science",
  "use": "literature-knowledge",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Ramprasad-Group/SALSA"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "gpt-4.1-mini",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-4.1-mini",
   "newest_release": "2025-04-14",
   "newest_class": "frontier-api",
   "lag_days": 505,
   "frontier_at_paper": "Claude Fable 5.1 (2026-09-01)",
   "flagship_releases_since": 32
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 24
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "general",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2609.24470",
  "date": "2026-09-21",
  "title": "Spatial Action Review: A Visual Analytics Dashboard for Auditing Language-to-Action Hand-offs in Electron Microscopy",
  "scope_group": "other",
  "use": "data-interpretation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Mohinta2892/SpatialActionReview"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen3-VL",
    "version": null,
    "role": "system"
   }
  ],
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 43
  }
 },
 {
  "id": "2609.25000",
  "date": "2026-09-21",
  "title": "Don't mind the gap: Absolutely Continuous Edge Spectrum in a Mobility Gap",
  "scope_group": "condensed-matter",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "none",
  "harness_class": "none",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [],
  "model_lag": {
   "newest_model": "gpt-6",
   "newest_release": "2026-09-04",
   "newest_class": "frontier-api",
   "lag_days": 17,
   "frontier_at_paper": "Grok 4.7 (2026-09-21)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 25
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2609.25384",
  "date": "2026-09-21",
  "title": "Lattice in Line: Optimized DMRG ordering for complex lattice geometries",
  "scope_group": "condensed-matter",
  "use": "code-algorithm-generation",
  "kind": "method",
  "harness": "none",
  "harness_class": "none",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/spinflip/lattice_in_line",
   "https://github.com/pdobsan/pynauty",
   "https://github.com/rliang/libtsqubo"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Fable 5",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "claude-fable-5",
   "newest_release": "2026-06-09",
   "newest_class": "frontier-api",
   "lag_days": 104,
   "frontier_at_paper": "Grok 4.7 (2026-09-21)",
   "flagship_releases_since": 7
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 35
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-materials",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-materials",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2609.27036",
  "date": "2026-09-22",
  "title": "An open benchmark for machine learning-based polymer property prediction",
  "scope_group": "materials-science",
  "use": "property-prediction",
  "kind": "benchmark",
  "harness": "none",
  "harness_class": "none",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated; polyBERT downloaded November 6th, 2025",
  "code_urls": [
   "https://github.com/rlearsch/PolymerBenchmark2026"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-4o",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "GPT-o3-mini",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "polyBERT",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "o3-mini",
   "newest_release": "2025-01-31",
   "newest_class": "frontier-api",
   "lag_days": 599,
   "frontier_at_paper": "Claude Opus 5.5 (2026-09-22)",
   "flagship_releases_since": 41
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "molecular-soft",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "molecular-soft",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2609.27302",
  "date": "2026-09-23",
  "title": "Unmodeled states and uncertain action outcomes in agentic scanning tunneling microscopy",
  "scope_group": "condensed-matter",
  "use": "experimental-control",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/BB-84C/quailbot-pi",
   "https://github.com/BB-84C/quail-cli-core",
   "https://github.com/BB-84C/NanonisSPMController-CLI",
   "https://github.com/earendil-works/pi",
   "https://pypi.org/project/nanonis-spm/"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "gpt-5.6-sol",
    "version": null,
    "role": "system"
   },
   {
    "name": "claude-opus-5",
    "version": null,
    "role": "system"
   },
   {
    "name": "gpt-5.4-mini",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-5",
   "newest_release": "2026-07-24",
   "newest_class": "frontier-api",
   "lag_days": 61,
   "frontier_at_paper": "Claude Opus 5.5 (2026-09-22)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 3,
   "cited_by_in_set": 0,
   "bibliography_size": 40
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "general",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2609.29492",
  "date": "2026-08-25",
  "title": "QEVOLVE-Bench: A Seed Benchmark for Quantum SDK Evolution and Repair Planning",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/FraQTech/qevolve"
  ],
  "ai_use_statement": false,
  "models": [],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 15
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2609.29595",
  "date": "2026-08-27",
  "title": "Building Atomic Structures from Natural Language",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "language model",
    "version": null,
    "role": "system"
   },
   {
    "name": "language model",
    "version": null,
    "role": "judge"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 21
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "general",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2609.30133",
  "date": "2026-09-24",
  "title": "AI-guided high-throughput discovery of iridium- and ruthenium-free palladium-oxide catalysts for durable acidic oxygen evolution",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/fl97inc/lila_oer_manuscript"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "LLMs",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Opus 4.6",
    "version": "4.6",
    "role": "baseline"
   },
   {
    "name": "Claude Opus 4.6",
    "version": "4.6",
    "role": "writing"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-4.6",
   "newest_release": "2026-02-05",
   "newest_class": "frontier-api",
   "lag_days": 231,
   "frontier_at_paper": "Claude Opus 5.5 (2026-09-22)",
   "flagship_releases_since": 17
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "catalysts",
    "question": "targeted-design"
   },
   "opus": {
    "subject": "catalysts",
    "question": "targeted-design"
   }
  }
 },
 {
  "id": "2609.30157",
  "date": "2026-09-24",
  "title": "Learning and interpreting policies for simultaneous entanglement requests in quantum networks",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "application",
  "harness": "none",
  "harness_class": "none",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "September 2026 availability stated; evaluation dates not stated",
  "code_urls": [
   "https://github.com/leonrode/simultaneous-entanglement-requests-in-quantum-networks"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Gemini 3.1 Pro",
    "version": "3.1 Pro",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gemini-3.1-pro",
   "newest_release": "2026-02-19",
   "newest_class": "frontier-api",
   "lag_days": 217,
   "frontier_at_paper": "Claude Opus 5.5 (2026-09-22)",
   "flagship_releases_since": 14
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 26
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-hardware",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-hardware",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2609.31679",
  "date": "2026-09-16",
  "title": "Toward AI-Assisted Poultry Coccidiosis Diagnosis: Evaluating Gemini and BiomedParse on Eimeria Microscopy Images",
  "scope_group": "other",
  "use": "data-interpretation",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Gemini",
    "version": "1.5",
    "role": "system"
   },
   {
    "name": "BiomedParse",
    "version": null,
    "role": "other"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 29
  }
 },
 {
  "id": "2609.31790",
  "date": "2026-09-25",
  "title": "CP-Agent: A Harness-Engineered Agent for Crystal Plasticity Simulation Workflows",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/samoalfred/harness-cp-agents"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT-4o",
    "version": null,
    "role": "system"
   },
   {
    "name": "ChatGPT",
    "version": null,
    "role": "writing"
   },
   {
    "name": "Claude",
    "version": null,
    "role": "writing"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-4o",
   "newest_release": "2024-05-13",
   "newest_class": "frontier-api",
   "lag_days": 865,
   "frontier_at_paper": "Claude Opus 5.5 (2026-09-22)",
   "flagship_releases_since": 41
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 52
  },
  "subject_codes": {
   "sol": {
    "subject": "alloys",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "alloys",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2609.33192",
  "date": "2026-09-27",
  "title": "AG-CoT: Verified Algorithmic Traces for LLM Program Synthesis on Clifford Circuits",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Qwen2.5-3B-Instruct",
    "version": null,
    "role": "system"
   },
   {
    "name": "Mistral-7B-Instruct-v0.3",
    "version": "c170c708c41dac9275d15a8fff4eca08d52bab71",
    "role": "system"
   },
   {
    "name": "Qwen2.5-Coder-32B-Instruct",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "qwen2.5-coder-32b",
   "newest_release": "2024-11-12",
   "newest_class": "open-weights",
   "lag_days": 684,
   "frontier_at_paper": "Claude Opus 5.5 (2026-09-22)",
   "flagship_releases_since": 41
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 40
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2609.34119",
  "date": "2026-09-28",
  "title": "The equivalence of topological indices for $\\mathrm{U}(1)\\rtimes\\mathbb{Z}_2$-symmetric quantum spin chains",
  "scope_group": "condensed-matter",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "none",
  "harness_class": "none",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [],
  "model_lag": {
   "newest_model": "gpt-6-astra-pro",
   "newest_release": "2026-09-04",
   "newest_class": "frontier-api",
   "lag_days": 24,
   "frontier_at_paper": "Claude Opus 5.5 (2026-09-22)",
   "flagship_releases_since": 2
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 42
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2609.34293",
  "date": "2026-09-28",
  "title": "Optimal Quantum Algorithms for Ordered Search",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Fable 5",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.6-Sol",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-6-Astra",
    "version": null,
    "role": "writing"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-sol",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 81,
   "frontier_at_paper": "Claude Opus 5.5 (2026-09-22)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 14
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "targeted-design"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2609.35581",
  "date": "2026-09-28",
  "title": "QC-Stark: A Multi-Task Benchmark Revealing Capability Dissociations in LLMs Evaluated on Quantum Computing Tasks",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 12,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://huggingface.co/datasets/pranavgupta/qc-stark"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "o4-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Sonnet 5",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemma-4 31B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Opus 4.1",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-4.1-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini Flash Lite",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini 3.5 Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "Mistral Large 3",
    "version": null,
    "role": "system"
   },
   {
    "name": "LLaMA-3.3 70B",
    "version": null,
    "role": "system"
   },
   {
    "name": "o3-mini",
    "version": null,
    "role": "system"
   },
   {
    "name": "o3",
    "version": null,
    "role": "system"
   },
   {
    "name": "Opus 4.6 and higher",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-5",
   "newest_release": "2026-06-30",
   "newest_class": "frontier-api",
   "lag_days": 90,
   "frontier_at_paper": "Claude Opus 5.5 (2026-09-22)",
   "flagship_releases_since": 8
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2609.36469",
  "date": "2026-09-29",
  "title": "From Automated Simulation to Autonomous Discovery: A Hierarchical Framework for Agentic Computational Materials Science",
  "scope_group": "materials-science",
  "use": "research-orchestration",
  "kind": "review-perspective",
  "harness": "none",
  "harness_class": "none",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "ai-workflow-perspective"
   },
   "opus": {
    "subject": "general",
    "question": "ai-workflow-perspective"
   }
  }
 },
 {
  "id": "2609.37053",
  "date": "2026-09-29",
  "title": "MatToolBench: Benchmarking Multimodal Agents in Real-World Materials Science Workflows",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "benchmark",
  "harness": "off-the-shelf: NaviAgent; custom runner/evaluators",
  "harness_class": "off-the-shelf",
  "base_models": 7,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/meiwu5/MatToolBench"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Doubao-seed-1-8",
    "version": null,
    "role": "system"
   },
   {
    "name": "Kimi-k2.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude-sonnet-4.6",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-VL-235B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-VL-32B",
    "version": null,
    "role": "system"
   },
   {
    "name": "Qwen3-VL-8B",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Doubao-seed-1-8",
    "version": null,
    "role": "judge"
   },
   {
    "name": "Gemini-3.1-pro",
    "version": "Gemini-3.1-pro-preview",
    "role": "judge"
   },
   {
    "name": "Qwen3-VL-235B",
    "version": null,
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4",
   "newest_release": "2026-03-05",
   "newest_class": "frontier-api",
   "lag_days": 208,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 14
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 37
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2609.38340",
  "date": "2026-09-29",
  "title": "CARAT: Do Materials LLMs Reason or Recite?",
  "scope_group": "materials-science",
  "use": "scientific-qa",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated; dated artifact names are provided",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Qwen/Qwen3.6-27B",
    "version": null,
    "role": "system"
   },
   {
    "name": "unnamed LLM judges",
    "version": null,
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "qwen3.6-27b",
   "newest_release": "2026-04-21",
   "newest_class": "open-weights",
   "lag_days": 161,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 13
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 28
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2609.39302",
  "date": "2026-09-30",
  "title": "SQD-Agent: LLM-driven agentic framework for Quantum Chemistry workflows",
  "scope_group": "other",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "off-the-shelf: OpenCode",
  "harness_class": "off-the-shelf",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/qiskit-community/fulqrum",
   "https://github.com/anomalyco/opencode"
  ],
  "ai_use_statement": true,
  "models": [],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 92
  }
 },
 {
  "id": "2610.00636",
  "date": "2026-09-30",
  "title": "CompMat-Bench: Benchmarking AI Agents for Computational Materials Science",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "benchmark",
  "harness": "off-the-shelf: MatClaw; Codex CLI 0.154.0",
  "harness_class": "off-the-shelf",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/cz2014/CompMat-Bench",
   "https://github.com/openai/codex"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "DeepSeek V4.1 Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini 3.8 Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.6 Sol",
    "version": null,
    "role": "system"
   },
   {
    "name": "unnamed language model",
    "version": null,
    "role": "data-generation"
   }
  ],
  "model_lag": {
   "newest_model": "deepseek-v4.1-flash",
   "newest_release": "2026-09-10",
   "newest_class": "open-weights",
   "lag_days": 20,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 4,
   "cited_by_in_set": 0,
   "bibliography_size": 61
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "general",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2610.00943",
  "date": "2026-10-01",
  "title": "Automated Many-Body Simulations of Strongly Correlated Systems Using a Correlation-Aware Agentic Framework",
  "scope_group": "condensed-matter",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/lith0419/CAFES"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "DeepSeek-v4.1-Flash",
    "version": "DeepSeek-v4.1-Flash",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "deepseek-v4.1-flash",
   "newest_release": "2026-09-10",
   "newest_class": "open-weights",
   "lag_days": 21,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 62
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-materials",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "quantum-materials",
    "question": "workflow-automation"
   }
  }
 },
 {
  "id": "2610.02671",
  "date": "2026-10-02",
  "title": "LATHE: LAnguage-driven Toolkit for Hypothesis-based crystal Editing",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "off-the-shelf: Pydantic AI; custom FastMCP workflow; Claude Code baseline",
  "harness_class": "off-the-shelf",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Fung-Lab/StructGradDesign"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "gpt-5.4-nano",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4-nano",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "gpt-4o",
    "version": null,
    "role": "system"
   },
   {
    "name": "gpt-4o",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "claude-opus-4-6",
    "version": "claude-opus-4-6",
    "role": "baseline"
   },
   {
    "name": "unspecified parser/verifier LLMs",
    "version": null,
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4-nano",
   "newest_release": "2026-03-17",
   "newest_class": "frontier-api",
   "lag_days": 199,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 14
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 60
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2610.04134",
  "date": "2026-10-02",
  "title": "Agentic Resource Allocation for Batch Multi-Objective Bayesian Optimization in Autonomous Materials Discovery",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": true,
  "evaluation_dates": "24–29 June 2026",
  "code_urls": [
   "https://github.com/Tripp5118/resource-allocation"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT-4o",
    "version": "gpt-4o",
    "role": "system"
   },
   {
    "name": "ChatGPT-5",
    "version": null,
    "role": "writing"
   },
   {
    "name": "Claude Sonnet-4",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-4",
   "newest_release": "2025-05-22",
   "newest_class": "frontier-api",
   "lag_days": 498,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 33
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 57
  },
  "subject_codes": {
   "sol": {
    "subject": "alloys",
    "question": "method-capability"
   },
   "opus": {
    "subject": "alloys",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2610.04866",
  "date": "2026-10-04",
  "title": "Phenomenon-first problem formulation with language models in metal-organic frameworks",
  "scope_group": "materials-science",
  "use": "reasoning-proof",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "MC-015: 28 September 2026; others not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT-6 Astra",
    "version": "gpt-6-astra",
    "role": "system"
   },
   {
    "name": "GPT-5.6 Sol",
    "version": null,
    "role": "other"
   },
   {
    "name": "gpt-5.5-2026-04-23",
    "version": "gpt-5.5-2026-04-23",
    "role": "other"
   },
   {
    "name": "Claude Opus 5",
    "version": null,
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-6-astra",
   "newest_release": "2026-09-04",
   "newest_class": "frontier-api",
   "lag_days": 30,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "porous-frameworks",
    "question": "method-capability"
   },
   "opus": {
    "subject": "porous-frameworks",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2610.05304",
  "date": "2026-10-04",
  "title": "Demonstration of a Structured Agentic Workflow for Applied Quantum Computing Research",
  "scope_group": "other",
  "use": "research-orchestration",
  "kind": "method",
  "harness": "custom QAOS with off-the-shelf: Claude Code",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Sonnet 5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Claude Sonnet 5",
    "version": null,
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "claude-sonnet-5",
   "newest_release": "2026-06-30",
   "newest_class": "frontier-api",
   "lag_days": 96,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 9
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 14
  }
 },
 {
  "id": "2610.06623",
  "date": "2026-10-05",
  "title": "Large Language Model-Guided Discovery of Weight-Five Bivariate Bicycle Codes",
  "scope_group": "adjacent-quantum",
  "use": "design-optimization",
  "kind": "application",
  "harness": "off-the-shelf: OpenEvolve 0.2.26",
  "harness_class": "off-the-shelf",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/qiskit-community/qcode-discovery",
   "https://github.com/QEC-pages/2BGA-codes",
   "https://github.com/algorithmicsuperintelligence/openevolve",
   "https://unitaryfoundation.github.io/qldpc-challenge/"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Opus 5",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.6 Sol",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-5",
   "newest_release": "2026-07-24",
   "newest_class": "frontier-api",
   "lag_days": 73,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 6
  },
  "citations": {
   "cites_in_set": 3,
   "cited_by_in_set": 0,
   "bibliography_size": 54
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "targeted-design"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "targeted-design"
   }
  }
 },
 {
  "id": "2610.07862",
  "date": "2026-10-06",
  "title": "A self-learning scientific agent for X-ray diffraction",
  "scope_group": "materials-science",
  "use": "data-interpretation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/DeltaGanjiang/deltaxrdbench",
   "https://github.com/DeltaGanjiang"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "LLM",
    "version": null,
    "role": "system"
   },
   {
    "name": "AutoXRD",
    "version": null,
    "role": "baseline"
   }
  ],
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 0,
   "bibliography_size": 68
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2610.10258",
  "date": "2026-10-07",
  "title": "QuSema: Detecting Silent Bugs in Quantum Libraries via Quantum-knowledge-enhanced Agents",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "DeepSeek-V4.1-Flash",
    "version": "V4.1-Flash",
    "role": "system"
   },
   {
    "name": "Opus 5",
    "version": "5",
    "role": "system"
   },
   {
    "name": "Fable 5",
    "version": "5",
    "role": "baseline"
   },
   {
    "name": "GPT-5.6-Sol",
    "version": "5.6-Sol",
    "role": "baseline"
   },
   {
    "name": "GPT-5.6-Sol",
    "version": "5.6-Sol",
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "deepseek-v4.1-flash",
   "newest_release": "2026-09-10",
   "newest_class": "open-weights",
   "lag_days": 27,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 3
  },
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 0,
   "bibliography_size": 136
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "method-capability"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2610.10879",
  "date": "2026-10-07",
  "title": "MD-LLM-2: A Transferable Language Model of Molecular Dynamics with Physical Conditioning and Explicit Path Probabilities",
  "scope_group": "other",
  "use": "property-prediction",
  "kind": "method",
  "harness": "none",
  "harness_class": "none",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Qwen2.5-14B",
    "version": null,
    "role": "system"
   },
   {
    "name": "ESM3",
    "version": "esm3_sm_open_v1",
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "qwen2.5-14b",
   "newest_release": "2024-09-19",
   "newest_class": "open-weights",
   "lag_days": 748,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 42
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 43
  }
 },
 {
  "id": "2610.12187",
  "date": "2026-10-08",
  "title": "Autonomous Code Migration for Quantum Programming Languages: A Case Study with QED-C Benchmarks",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "application",
  "harness": "off-the-shelf: Claude Code 2.1.156; Codex 0.131.0; Cursor Agent CLI 2026.05.28-a70ca7c",
  "harness_class": "off-the-shelf",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/SRI-International/QC-App-Oriented-Benchmarks",
   "https://github.com/NVIDIA/skills/tree/main/skills",
   "https://github.com/NVIDIA/cuda-quantum",
   "https://github.com/NVIDIA/cuda-q-academic"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Opus 4.7",
    "version": "4.7",
    "role": "system"
   },
   {
    "name": "GPT 5.5",
    "version": "5.5",
    "role": "system"
   },
   {
    "name": "Gemini 3.1 Pro",
    "version": "3.1 Pro",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 168,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 12
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 49
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2610.12700",
  "date": "2026-10-08",
  "title": "Improving the lower bound on the Laughlin gap: An SDP search",
  "scope_group": "condensed-matter",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://arxiv.org/src/2610.12700v1/anc/"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Opus 5.5",
    "version": "5.5",
    "role": "system"
   },
   {
    "name": "ChatGPT 6.1 Sol",
    "version": "6.1 Sol",
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-6.1-sol",
   "newest_release": "2026-09-29",
   "newest_class": "frontier-api",
   "lag_days": 9,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 0
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 15
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2610.13573",
  "date": "2026-10-09",
  "title": "MatQuery: Execution-Grounded Natural-Language Querying over Heterogeneous Materials Databases",
  "scope_group": "materials-science",
  "use": "literature-knowledge",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": false,
  "models": [
   {
    "name": "deepseek-flash",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "not stated",
    "version": null,
    "role": "system"
   },
   {
    "name": "not stated",
    "version": null,
    "role": "data-generation"
   },
   {
    "name": "not stated",
    "version": null,
    "role": "judge"
   },
   {
    "name": "not stated",
    "version": null,
    "role": "other"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 20
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "general",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2610.13811",
  "date": "2026-10-09",
  "title": "QTestBench: A Test Framework for Evaluating LLM-Generated Quantum Code",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 5,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated; model references accessed June 2026",
  "code_urls": [
   "https://github.com/eBrain4Everyone/QTestBench"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Opus 4.6",
    "version": "4.6",
    "role": "system"
   },
   {
    "name": "DeepSeek V3.2",
    "version": "V3.2",
    "role": "system"
   },
   {
    "name": "Gemini 3 Pro",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT-5.4",
    "version": "5.4",
    "role": "system"
   },
   {
    "name": "Qwen3-Coder",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4",
   "newest_release": "2026-03-05",
   "newest_class": "frontier-api",
   "lag_days": 218,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 14
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 29
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2610.16733",
  "date": "2026-10-09",
  "title": "TheoPhysEval: Assessing AI agents at autonomous theoretical physics problem solving",
  "scope_group": "other",
  "use": "reasoning-proof",
  "kind": "benchmark",
  "harness": "off-the-shelf: Harbor; Codex, Claude Code CLI, Antigravity",
  "harness_class": "off-the-shelf",
  "base_models": 5,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://huggingface.co/datasets/RameshChandra1/TheoPhysEval"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "GPT-6-Astra",
    "version": null,
    "role": "system"
   },
   {
    "name": "Opus 5.5",
    "version": null,
    "role": "system"
   },
   {
    "name": "Gemini 3.8 Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "DeepSeek-V4.1-Flash",
    "version": null,
    "role": "system"
   },
   {
    "name": "GLM 5.3",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-5.5",
   "newest_release": "2026-09-22",
   "newest_class": "frontier-api",
   "lag_days": 17,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 61
  }
 },
 {
  "id": "2610.17193",
  "date": "2026-10-12",
  "title": "Principles and Techniques for Agentic Research in Experimental Physics",
  "scope_group": "other",
  "use": "research-orchestration",
  "kind": "method",
  "harness": "off-the-shelf: Claude Code and Codex; custom PhysKit",
  "harness_class": "off-the-shelf",
  "base_models": 9,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Walsworth-Group/physkit",
   "https://github.com/karpathy/autoresearch"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Claude Opus 4.8",
    "version": "4.8",
    "role": "system"
   },
   {
    "name": "GPT 6 Astra",
    "version": "6 Astra",
    "role": "system"
   },
   {
    "name": "Claude Opus 5",
    "version": "5",
    "role": "system"
   },
   {
    "name": "Claude Fable 5",
    "version": "5",
    "role": "system"
   },
   {
    "name": "ChatGPT 5.6 Sol",
    "version": "5.6 Sol",
    "role": "system"
   },
   {
    "name": "Claude Fable 5.1",
    "version": "5.1",
    "role": "system"
   },
   {
    "name": "Opus 5.5",
    "version": "5.5",
    "role": "system"
   },
   {
    "name": "Claude Sonnet 4.5",
    "version": "4.5",
    "role": "system"
   },
   {
    "name": "Sonnet 5",
    "version": "5",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-opus-5.5",
   "newest_release": "2026-09-22",
   "newest_class": "frontier-api",
   "lag_days": 20,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 1
  },
  "citations": {
   "cites_in_set": 2,
   "cited_by_in_set": 0,
   "bibliography_size": 145
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "ai-workflow-perspective"
   },
   "opus": {
    "subject": "general",
    "question": "ai-workflow-perspective"
   }
  }
 },
 {
  "id": "2610.17286",
  "date": "2026-10-12",
  "title": "Quantum Cryptography and Autoformalization in Lean",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 3,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://maorbenshahar.github.io/QCryptLean/docs/",
   "https://github.com/maorbenshahar/loopy-prover",
   "https://github.com/maorbenshahar/lean-graph-tools"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Opus 5",
    "version": "Opus 5",
    "role": "system"
   },
   {
    "name": "Claude",
    "version": null,
    "role": "judge"
   },
   {
    "name": "GPT",
    "version": null,
    "role": "system"
   },
   {
    "name": "GPT",
    "version": null,
    "role": "judge"
   },
   {
    "name": "GLM-5.3-Flash",
    "version": "GLM-5.3-Flash",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "glm-5.3-flash",
   "newest_release": "2026-08-26",
   "newest_class": "open-weights",
   "lag_days": 47,
   "frontier_at_paper": "GPT-6.1 Sol (2026-09-29)",
   "flagship_releases_since": 5
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 48
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   },
   "opus": {
    "subject": "quantum-theory",
    "question": "physical-findings"
   }
  }
 },
 {
  "id": "2610.17708",
  "date": "2026-10-13",
  "title": "QEDIA: Quantum Program Verification in Lean",
  "scope_group": "adjacent-quantum",
  "use": "reasoning-proof",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/qedia/quantum-formal-verification"
  ],
  "ai_use_statement": true,
  "models": [],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 52
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2610.17804",
  "date": "2026-10-13",
  "title": "Learning from fracture: Teaching language models to design through physical experience",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/GenerativeMechanicsLab/FracLearn"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "gpt-5.4",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.4",
   "newest_release": "2026-03-05",
   "newest_class": "frontier-api",
   "lag_days": 222,
   "frontier_at_paper": "Gemini 4 Argon (2026-10-13)",
   "flagship_releases_since": 15
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 60
  },
  "subject_codes": {
   "sol": {
    "subject": "structural-composites",
    "question": "method-capability"
   },
   "opus": {
    "subject": "structural-composites",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2610.17832",
  "date": "2026-10-13",
  "title": "Agentic-MatriBO: Capability-Compiled Agentic Bayesian Optimization for Materials and Chemical Design",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "method",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 4,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/huawei-noah/HEBO",
   "https://github.com/Xinzhe309/LABO",
   "https://github.com/richardcsuwandi/cake",
   "https://github.com/giang-n-ngo/lmabo",
   "https://github.com/Ablatif6c/bora-the-explorer",
   "https://github.com/little1d/Reasoning-BO",
   "https://github.com/tennisonliu/LLAMBO"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "DeepSeek 4.0 Flash",
    "version": "DeepSeek-V4-Flash",
    "role": "system"
   },
   {
    "name": "DeepSeek 4.0 Pro",
    "version": "DeepSeek-V4-Pro",
    "role": "system"
   },
   {
    "name": "Qwen3-Max",
    "version": "qwen3-max",
    "role": "system"
   },
   {
    "name": "GPT-5.6-sol",
    "version": "GPT-5.6-sol",
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6-sol",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 96,
   "frontier_at_paper": "Gemini 4 Argon (2026-10-13)",
   "flagship_releases_since": 8
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 32
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "method-capability"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2610.18949",
  "date": "2026-10-13",
  "title": "Virtual-Alloy Physics-Aware Semantic System for extrapolative prediction of perovskite solar cells",
  "scope_group": "materials-science",
  "use": "property-prediction",
  "kind": "method",
  "harness": "none",
  "harness_class": "none",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [],
  "ai_use_statement": true,
  "models": [
   {
    "name": "SciBERT",
    "version": null,
    "role": "system"
   },
   {
    "name": "SciBERT",
    "version": null,
    "role": "baseline"
   },
   {
    "name": "Claude",
    "version": null,
    "role": "writing"
   },
   {
    "name": "Gemini",
    "version": null,
    "role": "writing"
   }
  ],
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "semiconductor-optical",
    "question": "method-capability"
   },
   "opus": {
    "subject": "semiconductor-optical",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2610.19404",
  "date": "2026-10-14",
  "title": "ActMOF: Mapping Reticular Synthesis through Active Learning",
  "scope_group": "materials-science",
  "use": "design-optimization",
  "kind": "benchmark",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 0,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "August 2026 snapshots; evaluation dates not stated",
  "code_urls": [
   "https://github.com/deep-synthesis/ActMOF"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "gpt-4o",
    "version": "gpt-4o-2024-08-06",
    "role": "baseline"
   },
   {
    "name": "gpt-5.5",
    "version": "gpt-5.5-2026-04-23",
    "role": "baseline"
   },
   {
    "name": "gpt-5-mini",
    "version": "gpt-5-mini-2025-08-07",
    "role": "baseline"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 174,
   "frontier_at_paper": "Gemini 4 Argon (2026-10-13)",
   "flagship_releases_since": 13
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 112
  },
  "subject_codes": {
   "sol": {
    "subject": "porous-frameworks",
    "question": "model-evaluation"
   },
   "opus": {
    "subject": "porous-frameworks",
    "question": "model-evaluation"
   }
  }
 },
 {
  "id": "2610.20779",
  "date": "2026-10-14",
  "title": "Enabling large language model agents for high-fidelity many-body perturbation theory calculations",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "custom; off-the-shelf: Claude Code, Codex",
  "harness_class": "off-the-shelf",
  "base_models": 7,
  "dated_system_snapshot": true,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/Li-Group-USC/AI-Donkey-Release.git",
   "https://github.com/Li-Group-USC/AI-Donkey-Paper-Artifacts-Release.git",
   "https://www.abinit.org",
   "http://www.wannier.org",
   "https://berkeleygw.org",
   "https://www.quantum-espresso.org",
   "https://docs.epw-code.org"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "Haiku 4.5",
    "version": "claude-haiku-4-5-20251001",
    "role": "system"
   },
   {
    "name": "Sonnet 4.6",
    "version": "claude-sonnet-4-6",
    "role": "system"
   },
   {
    "name": "Opus 4.7",
    "version": "claude-opus-4-7",
    "role": "system"
   },
   {
    "name": "GPT-5.4-mini",
    "version": "gpt-5.4-mini",
    "role": "system"
   },
   {
    "name": "GPT-5.5",
    "version": "gpt-5.5",
    "role": "system"
   },
   {
    "name": "gpt-oss-120b",
    "version": "openai/gpt-oss-120b",
    "role": "system"
   },
   {
    "name": "Gemma-4-31B",
    "version": "google/gemma-4-31B-it",
    "role": "system"
   },
   {
    "name": "unspecified LLM judge",
    "version": null,
    "role": "judge"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.5",
   "newest_release": "2026-04-23",
   "newest_class": "frontier-api",
   "lag_days": 174,
   "frontier_at_paper": "Gemini 4 Argon (2026-10-13)",
   "flagship_releases_since": 13
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 1
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "workflow-automation"
   },
   "opus": {
    "subject": "general",
    "question": "method-capability"
   }
  }
 },
 {
  "id": "2610.21015",
  "date": "2026-10-14",
  "title": "MASON: Materials Atomistic Structure builder Operated in Natural language",
  "scope_group": "materials-science",
  "use": "simulation-workflow",
  "kind": "method",
  "harness": "unclear",
  "harness_class": "unclear",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/zxy-boop/MASON"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "GPT-5.6",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "gpt-5.6",
   "newest_release": "2026-07-09",
   "newest_class": "frontier-api",
   "lag_days": 97,
   "frontier_at_paper": "Gemini 4 Argon (2026-10-13)",
   "flagship_releases_since": 8
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 43
  },
  "subject_codes": {
   "sol": {
    "subject": "general",
    "question": "research-infrastructure"
   },
   "opus": {
    "subject": "general",
    "question": "research-infrastructure"
   }
  }
 },
 {
  "id": "2610.21050",
  "date": "2026-10-15",
  "title": "Large Language Model-Assisted Systematic Study of Anomalous Adsorption Phenomena",
  "scope_group": "materials-science",
  "use": "literature-knowledge",
  "kind": "application",
  "harness": "custom",
  "harness_class": "custom",
  "base_models": 2,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://doi.org/10.5281/zenodo.23381050"
  ],
  "ai_use_statement": true,
  "models": [
   {
    "name": "glm-4.5-air",
    "version": "glm-4.5-air",
    "role": "system"
   },
   {
    "name": "gemini-2.5-pro",
    "version": "gemini-2.5-pro",
    "role": "system"
   },
   {
    "name": "Qwen3-235B-A22B",
    "version": "Qwen3-235B-A22B",
    "role": "baseline"
   },
   {
    "name": "Claude 5",
    "version": null,
    "role": "writing"
   },
   {
    "name": "Nano Banana 2",
    "version": null,
    "role": "other"
   }
  ],
  "model_lag": {
   "newest_model": "glm-4.5-air",
   "newest_release": "2025-07-28",
   "newest_class": "open-weights",
   "lag_days": 444,
   "frontier_at_paper": "Gemini 4 Argon (2026-10-13)",
   "flagship_releases_since": 31
  },
  "citations": {
   "cites_in_set": 0,
   "cited_by_in_set": 0,
   "bibliography_size": 46
  },
  "subject_codes": {
   "sol": {
    "subject": "porous-frameworks",
    "question": "literature-findings"
   },
   "opus": {
    "subject": "porous-frameworks",
    "question": "literature-findings"
   }
  }
 },
 {
  "id": "2610.22367",
  "date": "2026-10-15",
  "title": "Discovering state-of-the-art quantum circuits with large language models",
  "scope_group": "adjacent-quantum",
  "use": "code-algorithm-generation",
  "kind": "method",
  "harness": "off-the-shelf: OpenEvolve",
  "harness_class": "off-the-shelf",
  "base_models": 1,
  "dated_system_snapshot": false,
  "eval_dates_stated": false,
  "evaluation_dates": "not stated",
  "code_urls": [
   "https://github.com/andigu/uncompute",
   "https://github.com/algorithmicsuperintelligence/openevolve",
   "https://github.com/sipa/safegcd-bounds",
   "https://github.com/bitcoin-core/secp256k1/blob/master/doc/safegcd_implementation.md"
  ],
  "ai_use_statement": false,
  "models": [
   {
    "name": "Claude Mythos 5",
    "version": null,
    "role": "system"
   }
  ],
  "model_lag": {
   "newest_model": "claude-mythos-5",
   "newest_release": "2026-06-09",
   "newest_class": "frontier-api",
   "lag_days": 128,
   "frontier_at_paper": "Gemini 4 Argon (2026-10-13)",
   "flagship_releases_since": 10
  },
  "citations": {
   "cites_in_set": 1,
   "cited_by_in_set": 0,
   "bibliography_size": 76
  },
  "subject_codes": {
   "sol": {
    "subject": "quantum-software",
    "question": "targeted-design"
   },
   "opus": {
    "subject": "quantum-software",
    "question": "targeted-design"
   }
  }
 }
]