{ "generated": "2026-11-14", "count": 8, "models": [ { "slug": "alphafold3", "name": "AlphaFold 3", "kind": "co-folding", "publisher": "Google DeepMind", "tier": "verified", "license_code": { "status": "permissive", "spdx": "Apache-2.0", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "repo-license", "excerpt": "Apache License", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/LICENSE", "snapshot_id": "s03", "char_start": 34, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "license_weights": { "status": "restricted", "spdx": "LicenseRef-Google-DeepMind-terms", "custom_license": true, "commercial_use": "no", "derivatives": "unstated", "redistribution": "no", "license_conditions": [ "registration-required", "territory-dependent", "notice-required", "attribution" ], "requires_application": false, "evidence": [ { "location": "terms-page", "excerpt": "The AlphaFold 3 model parameters and output are **only** available for", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/WEIGHTS_TERMS_OF_USE.md", "snapshot_id": "s12", "char_start": 821, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" }, { "location": "terms-page", "excerpt": "You ***must not* publish or share AlphaFold 3 model parameters**", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/WEIGHTS_TERMS_OF_USE.md", "snapshot_id": "s12", "char_start": 1752, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" }, { "location": "terms-page", "excerpt": "The AlphaFold 3 model parameters and output are **only** available for non-commercial use by, or on behalf of, non-commercial organizations (*i.e.*, universities, non-profit organizations and research institutes, educational, journalism and government bodies).", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/WEIGHTS_TERMS_OF_USE.md", "snapshot_id": "s08", "char_start": null, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" }, { "location": "model-card-page", "excerpt": "not (i) resident of a embargoed country, (ii) ordinarily resident in a US", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/WEIGHTS_TERMS_OF_USE.md", "snapshot_id": "s08", "char_start": null, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" }, { "location": "model-card-page", "excerpt": "Google may at any time suspend or terminate your right to use and, as applicable", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/WEIGHTS_TERMS_OF_USE.md", "snapshot_id": "s08", "char_start": null, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "last_verified": "2026-11-14", "schema_version": "2.0", "record_type": "model", "official_urls": { "repository_or_release": "https://github.com/google-deepmind/alphafold3", "weights_page": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/WEIGHTS_TERMS_OF_USE.md", "paper": "https://arxiv.org/abs/2510.24670", "license_file": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/LICENSE" }, "version_history": [ { "version": "unspecified", "released_on": "2026-06-09", "code_license": null, "weights_license": null, "changed": "other", "note": null, "evidence": [ { "location": "official-change-notice", "excerpt": "Only the code is re-licensed, the weights license and terms of use stay the same.", "url": "https://github.com/google-deepmind/alphafold3/commit/7b197fe859", "snapshot_id": "s13", "char_start": 2617, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Change the code license from CC BY-NC-SA 4.0 to Apache 2.0 (commit 7b197fe)", "authors_or_org": "Google DeepMind (committed by Augustin-Zidek)", "date": "2026-06-09", "date_as_written": "2026-06-09", "source_type": "official-change-notice", "open_access": true, "doi_or_url": "https://github.com/google-deepmind/alphafold3/commit/7b197fe859" } } ] }, { "version": "AlphaFold 3", "released_on": "2026-07-23", "code_license": null, "weights_license": null, "changed": "other", "note": null, "evidence": [ { "location": "official-change-notice", "excerpt": "request access to the AlphaFold 3 model parameters, please complete", "url": "https://github.com/google-deepmind/alphafold3/commit/dd1a7badb6", "snapshot_id": "s14", "char_start": 3607, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Update the instructions on how to obtain AlphaFold 3 weights (commit dd1a7ba, removed lines)", "authors_or_org": "Google DeepMind (committed by Augustin-Zidek)", "date": "2026-07-23", "date_as_written": "2026-07-23", "source_type": "official-change-notice", "open_access": true, "doi_or_url": "https://github.com/google-deepmind/alphafold3/commit/dd1a7badb6" } } ] } ], "output_restrictions": [ { "kind": "no-training-other-models-on-output", "applies_to": "weights", "statement": { "text": { "en": "Output may not be used to train or create machine learning models or related technology for biomolecular structure prediction similar to AlphaFold 3.", "zh": "不得使用输出结果训练或构建与 AlphaFold 3 类似的生物分子结构预测机器学习模型或相关技术。" }, "evidence": [ { "location": "terms-page", "excerpt": "Use Output to train or create machine learning models or related technology for biomolecular structure prediction similar to AlphaFold 3", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/WEIGHTS_PROHIBITED_USE_POLICY.md", "snapshot_id": "s05", "char_start": null, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } }, { "kind": "non-commercial-output", "applies_to": "weights", "statement": { "text": { "en": "The model parameters and output are available only for non-commercial use by or on behalf of non-commercial organizations such as universities, non-profits, research institutes, educational, journalism and government bodies.", "zh": "模型参数及其输出仅限大学、非营利组织、研究机构、教育机构、新闻机构和政府机构等非商业组织自身或代表其进行非商业使用。" }, "evidence": [ { "location": "terms-page", "excerpt": "The AlphaFold 3 model parameters and output are **only** available for non-commercial use by, or on behalf of, non-commercial organizations (*i.e.*, universities, non-profit organizations and research institutes, educational, journalism and government bodies).", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/WEIGHTS_TERMS_OF_USE.md", "snapshot_id": "s08", "char_start": null, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } }, { "kind": "no-clinical-use", "applies_to": "all", "statement": { "text": { "en": "AlphaFold 3 and its output are for theoretical modeling only and are not intended, validated, or approved for clinical use.", "zh": "AlphaFold 3 及其输出仅用于理论建模，并非为临床用途设计、验证或批准。" }, "evidence": [ { "location": "readme", "excerpt": "AlphaFold 3 and its output are for theoretical modeling only. They are not intended, validated, or approved for clinical use.", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/HEAD/README.md", "snapshot_id": "s02", "char_start": null, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } }, { "kind": "non-commercial-output", "applies_to": "all", "statement": { "text": { "en": "Output Terms of Use (last modified 2024-11-09) forbid using AF3 output on behalf of commercial organisations or in commercial activities, including research for them.", "zh": "《输出使用条款》（最后修改 2024-11-09）禁止代表商业机构或在商业活动中使用 AF3 输出，包括为其开展的研究。" }, "evidence": [ { "location": "dataset-page", "excerpt": "Output are made available free of charge, for non-commercial use", "url": "https://github.com/google-deepmind/alphafold3/blob/main/OUTPUT_TERMS_OF_USE.md", "snapshot_id": "s15", "char_start": 3076, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "AlphaFold 3 Output Terms of Use", "authors_or_org": "Google DeepMind", "date": "2024-11-09", "date_as_written": "2024-11-09", "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/google-deepmind/alphafold3/blob/main/OUTPUT_TERMS_OF_USE.md" } } ] } }, { "kind": "no-training-other-models-on-output", "applies_to": "all", "statement": { "text": { "en": "Output must not be used to train or create biomolecular structure prediction models similar to AF3, including via distillation.", "zh": "不得用 AF3 输出训练或构建与 AF3 类似的生物分子结构预测模型，包括通过蒸馏。" }, "evidence": [ { "location": "dataset-page", "excerpt": "biomolecular structure prediction similar to AlphaFold 3 as made available", "url": "https://github.com/google-deepmind/alphafold3/blob/main/OUTPUT_TERMS_OF_USE.md", "snapshot_id": "s15", "char_start": 5970, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "AlphaFold 3 Output Terms of Use", "authors_or_org": "Google DeepMind", "date": "2024-11-09", "date_as_written": "2024-11-09", "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/google-deepmind/alphafold3/blob/main/OUTPUT_TERMS_OF_USE.md" } } ] } }, { "kind": "non-commercial-output", "applies_to": "weights", "statement": { "text": { "en": "Source sentence: You must not share Output with any commercial organization.", "zh": "来源原句：You must not share Output with any commercial organization." }, "evidence": [ { "location": "model-card-page", "excerpt": "You must not share Output with any commercial organization.", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/OUTPUT_TERMS_OF_USE.md", "snapshot_id": "s15", "char_start": null, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } }, { "kind": "other", "applies_to": "weights", "statement": { "text": { "en": "Source sentence: Distribute Output, or disclose findings arising from using AlphaFold 3 without citing our paper:", "zh": "来源原句：Distribute Output, or disclose findings arising from using AlphaFold 3 without citing our paper:" }, "evidence": [ { "location": "model-card-page", "excerpt": "Distribute Output, or disclose findings arising from using AlphaFold 3 without citing our paper:", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/WEIGHTS_PROHIBITED_USE_POLICY.md", "snapshot_id": "s05", "char_start": null, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } }, { "kind": "other", "applies_to": "weights", "statement": { "text": { "en": "Source sentence: making decisions in domains that affect material or individual rights or", "zh": "来源原句：making decisions in domains that affect material or individual rights or" }, "evidence": [ { "location": "model-card-page", "excerpt": "making decisions in domains that affect material or individual rights or", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/OUTPUT_TERMS_OF_USE.md", "snapshot_id": "s15", "char_start": null, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } } ], "training_data": [ { "resource": "training data", "slug": null, "role": "other", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "No structural training data released after 30 September 2021 were used.", "zh": "训练所用结构数据均不晚于 2021-09-30 发布。" }, "evidence": [ { "location": "peer-reviewed-paper", "excerpt": "No structural data used during training were released after 30 September 2021", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11168924/", "snapshot_id": "s11", "char_start": 53237, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Accurate structure prediction of biomolecular interactions with AlphaFold 3", "authors_or_org": "Abramson J, Adler J, Dunger J, et al. (Nature)", "date": "2024-05-08", "date_as_written": "2024-05-08", "source_type": "peer-reviewed-paper", "open_access": true, "doi_or_url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11168924/" } } ] } }, { "resource": "training data", "slug": null, "role": "other", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "Training data were enriched with AlphaFold-Multimer v2.3 predicted structures (cross-distillation).", "zh": "训练数据中加入了 AlphaFold-Multimer v2.3 的预测结构（交叉蒸馏）。" }, "evidence": [ { "location": "peer-reviewed-paper", "excerpt": "we use a cross-distillation method in which we enrich the training data with structures predicted by AlphaFold-Multimer (v.2.3)", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11168924/", "snapshot_id": "s11", "char_start": 22274, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Accurate structure prediction of biomolecular interactions with AlphaFold 3", "authors_or_org": "Abramson J, Adler J, Dunger J, et al. (Nature)", "date": "2024-05-08", "date_as_written": "2024-05-08", "source_type": "peer-reviewed-paper", "open_access": true, "doi_or_url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11168924/" } } ] } } ], "benchmark_overlap": [ { "benchmark": "PoseBusters", "benchmark_slug": null, "finding": "cutoff-after-benchmark", "citation": { "reference": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11168924/", "statement": { "text": { "en": "Because the standard cutoff is in 2021 and PoseBusters contains structures released in 2021 or later, AF3 authors trained a separate earlier-cutoff model for PoseBusters evaluation.", "zh": "由于标准截止日期在 2021 年，而 PoseBusters 含 2021 年及以后发布的结构，AF3 作者为 PoseBusters 评测另训了一个截止更早的模型。" }, "evidence": [ { "location": "peer-reviewed-paper", "excerpt": "As our standard training cut-off date is in 2021, we trained a separate AF3 model with an earlier training-set cutoff ( Methods ).", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11168924/", "snapshot_id": "s11", "char_start": 28438, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Accurate structure prediction of biomolecular interactions with AlphaFold 3", "authors_or_org": "Abramson J, Adler J, Dunger J, et al. (Nature)", "date": "2024-05-08", "date_as_written": "2024-05-08", "source_type": "peer-reviewed-paper", "open_access": true, "doi_or_url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11168924/" } } ] } } }, { "benchmark": "benchmark", "benchmark_slug": null, "finding": "overlap-reported", "citation": { "reference": "https://pmc.ncbi.nlm.nih.gov/articles/PMC12501370/", "statement": { "text": { "en": "Masters, Mahmoud and Lill tested AF3, RoseTTAFold All-Atom, Chai-1 and Boltz-1 with adversarial binding-site edits: with all binding-site residues mutated to glycine, all four still predicted the same ATP–CDK2 binding mode, which the authors read as possible overfitting to training-data features.", "zh": "Masters、Mahmoud、Lill 用对抗性结合位点改造测试 AF3、RoseTTAFold All-Atom、Chai-1、Boltz-1：把结合位点残基全部突变为甘氨酸后，四个模型仍预测相同的 ATP–CDK2 结合模式，作者认为可能是对训练数据特征的过拟合。" }, "evidence": [ { "location": "peer-reviewed-paper", "excerpt": "All four co-folding models continue to predict the ATP-CDK2 complex with the same binding mode, despite the loss of all major side-chain interactions.", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC12501370/", "snapshot_id": "s18", "char_start": 15315, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Investigating whether deep learning models for co-folding learn the physics of protein-ligand interactions", "authors_or_org": "Masters MR, Mahmoud AH, Lill MA (Nature Communications)", "date": "2025-10-06", "date_as_written": "2025-10-06", "source_type": "peer-reviewed-paper", "open_access": true, "doi_or_url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC12501370/" } } ] } } }, { "benchmark": "Runs N' Poses", "benchmark_slug": null, "finding": "memorisation-reported", "citation": { "reference": "https://github.com/plinder-org/runs-n-poses", "statement": { "text": { "en": "Runs N' Poses is designed as a zero-shot benchmark for methods with a 30 September 2021 structural cutoff (AF3's cutoff); its authors conclude current co-folding methods largely memorise training ligand poses (the abstract does not name the four methods; the names were not verified).", "zh": "Runs N' Poses 是为结构截止日期为 2021-09-30（即 AF3 的截止日期）的方法设计的零样本基准；作者结论是现有共折叠方法很大程度上记忆训练集配体构象（摘要未列出所评四个方法的名称，未核实）。" }, "evidence": [ { "location": "dataset-page", "excerpt": "This is a zero-shot benchmark, provided that your method uses a structural training cutoff of 30 September 2021.", "url": "https://github.com/plinder-org/runs-n-poses", "snapshot_id": "s16", "char_start": 4002, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "plinder-org/runs-n-poses README", "authors_or_org": "Škrinjar P, Eberhardt J, Tauriello G, Schwede T, Durairaj J (PLINDER org)", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/plinder-org/runs-n-poses" } } ] } } } ], "access": [ { "method": "open-weights-download", "url": "https://storage.googleapis.com/alphafold3/af3.bin.zst", "terms_differ_from_weights_license": false, "own_terms": null }, { "method": "hosted-api", "url": "https://alphafoldserver.com", "terms_differ_from_weights_license": false, "own_terms": null }, { "method": "third-party-hosted", "url": "https://github.com/google-deepmind/alphafold3/blob/main/README.md", "terms_differ_from_weights_license": false, "own_terms": { "text": { "en": "AlphaFold Server (alphafoldserver.com) offers AF3 for non-commercial use with a more limited set of ligands and modifications; commercial use is offered through Google Cloud.", "zh": "AlphaFold Server（alphafoldserver.com）提供非商业用途的 AF3，配体与修饰种类较少；商业使用经由 Google Cloud 提供。" }, "evidence": [ { "location": "dataset-page", "excerpt": "for non-commercial use, though with a more limited set of ligands and covalent modifications.", "url": "https://github.com/google-deepmind/alphafold3/blob/main/README.md", "snapshot_id": "s10", "char_start": null, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "AlphaFold 3 README", "authors_or_org": "Google DeepMind", "date": "2026-08-21", "date_as_written": "2026-08-21 (Cloud links added)", "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/google-deepmind/alphafold3/blob/main/README.md" } } ] } } ], "conflicts": [], "hardware_notes": [ { "text": { "en": "Requires Linux, up to 1 TB disk for genetic databases and an NVIDIA GPU with Compute Capability 8.0+; inputs up to 5,120 tokens verified on a single A100 80 GB or H100 80 GB.", "zh": "需要 Linux、最多 1 TB 磁盘存放遗传数据库、计算能力 8.0 及以上的 NVIDIA GPU；已验证单张 A100 80GB 或 H100 80GB 可处理最多 5,120 个 token 的输入。" }, "evidence": [ { "location": "dataset-page", "excerpt": "We have verified that inputs with up to 5,120 tokens can fit on a single NVIDIA A100 80 GB, or a single NVIDIA H100 80 GB.", "url": "https://github.com/google-deepmind/alphafold3/blob/main/docs/installation.md", "snapshot_id": "s17", "char_start": null, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "AlphaFold 3 installation documentation", "authors_or_org": "Google DeepMind", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/google-deepmind/alphafold3/blob/main/docs/installation.md" } } ] } ], "known_defects": [], "fit": { "suited_for": [ { "text": { "en": "Proteins with shallow MSAs are predicted with lower accuracy (AF3 paper).", "zh": "MSA 较浅的蛋白预测精度较低（AF3 论文）。" }, "evidence": [ { "location": "peer-reviewed-paper", "excerpt": "proteins with shallow MSAs are predicted with lower accuracy", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11168924/", "snapshot_id": "s11", "char_start": 39846, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Accurate structure prediction of biomolecular interactions with AlphaFold 3", "authors_or_org": "Abramson J, Adler J, Dunger J, et al. (Nature)", "date": "2024-05-08", "date_as_written": "2024-05-08", "source_type": "peer-reviewed-paper", "open_access": true, "doi_or_url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC11168924/" } } ] } ], "not_suited_for": [] }, "gaps": [ { "field": "known_defects", "note": "未找到有出处的说明 / no sourced statement found" } ], "completeness": { "core_complete": true }, "changelog": [ { "date": "2026-11-14", "change": { "en": "Listed on 2026-11-14. Judged by public rules and checked a second time before listing. The two passes disagreed; the stricter reading is carried.", "zh": "2026-11-14 上架。按公开规则判定，经二次核对后收录。两遍判定不一致，卡上取较严的读法。" }, "by": "second-pass" } ], "license_review": { "method": "rules-plus-second-pass-model", "method_label": { "en": "Judged by public rules and checked a second time before listing.", "zh": "按公开规则判定，经二次核对后收录。" }, "readings": [ { "restatement": { "en": "First pass (rules).", "zh": "第一遍（规则）。" }, "quotes": [ { "excerpt": "The AlphaFold 3 model parameters and output are **only** available for", "snapshot_id": "s12", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/WEIGHTS_TERMS_OF_USE.md" }, { "excerpt": "You ***must not* publish or share AlphaFold 3 model parameters**", "snapshot_id": "s12", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/WEIGHTS_TERMS_OF_USE.md" }, { "excerpt": "The AlphaFold 3 model parameters and output are **only** available for non-commercial use by, or on behalf of, non-commercial organizations (*i.e.*, universities, non-profit organizations and research institutes, educational, journalism and government bodies).", "snapshot_id": "s08", "url": "https://raw.githubusercontent.com/google-deepmind/alphafold3/main/WEIGHTS_TERMS_OF_USE.md" } ], "values": { "license_status": "restricted", "commercial_use": "no", "derivatives": "unstated", "redistribution": "unstated" }, "source": "rules" }, { "restatement": { "en": "Second pass: weights.redistribution: 原文明确不得发布或分享模型参数（仅可在本组织内按条款分享），个人获授权者不得与任何人分享。第一遍把该句列为证据却填 unstated。应为 no。 | missed sentence (class A): You must not share Output with any commercial organization. | missed sentence (class D): Distribute Output, or disclose findings arising from using AlphaFold 3 | missed sentence (class A?): making decisions in domains that affect material or individual rights | missed sentence (class D): not (i) resident of a embargoed country, (ii) ordinarily resident in a | missed sentence (class A?): Google may at any time suspend or terminate your right to use and, as | missed sentence (class A): You may only use AlphaFold 3 model parameters if received directly fro", "zh": "第二遍复核：weights.redistribution: 原文明确不得发布或分享模型参数（仅可在本组织内按条款分享），个人获授权者不得与任何人分享。第一遍把该句列为证据却填 unstated。应为 no。；漏读句（A 类）：You must not share Output with any commercial organization.；漏读句（D 类）：Distribute Output, or disclose findings arising from using AlphaFold 3；漏读句（A? 类）：making decisions in domains that affect material or individual rights ；漏读句（D 类）：not (i) resident of a embargoed country, (ii) ordinarily resident in a；漏读句（A? 类）：Google may at any time suspend or terminate your right to use and, as ；漏读句（A 类）：You may only use AlphaFold 3 model parameters if received directly fro" }, "quotes": [ { "excerpt": "You must not share Output with any commercial organization.", "snapshot_id": "s15", "url": null }, { "excerpt": "Distribute Output, or disclose findings arising from using AlphaFold 3 without citing our paper:", "snapshot_id": "s05", "url": null }, { "excerpt": "making decisions in domains that affect material or individual rights or", "snapshot_id": "s15", "url": null }, { "excerpt": "not (i) resident of a embargoed country, (ii) ordinarily resident in a US", "snapshot_id": "s08", "url": null } ], "values": { "license_status": "restricted", "commercial_use": "no", "derivatives": "unstated", "redistribution": "no" }, "source": "second-pass" } ], "adopted": 1, "second_pass": { "reviewer": "model-second-pass", "reviewed_at": "2026-11-13", "overall": "stricter", "agreed": false, "disagreement": { "en": "Second pass: weights.redistribution: 原文明确不得发布或分享模型参数（仅可在本组织内按条款分享），个人获授权者不得与任何人分享。第一遍把该句列为证据却填 unstated。应为 no。 | missed sentence (class A): You must not share Output with any commercial organization. | missed sentence (class D): Distribute Output, or disclose findings arising from using AlphaFold 3 | missed sentence (class A?): making decisions in domains that affect material or individual rights | missed sentence (class D): not (i) resident of a embargoed country, (ii) ordinarily resident in a | missed sentence (class A?): Google may at any time suspend or terminate your right to use and, as | missed sentence (class A): You may only use AlphaFold 3 model parameters if received directly fro", "zh": "第二遍复核：weights.redistribution: 原文明确不得发布或分享模型参数（仅可在本组织内按条款分享），个人获授权者不得与任何人分享。第一遍把该句列为证据却填 unstated。应为 no。；漏读句（A 类）：You must not share Output with any commercial organization.；漏读句（D 类）：Distribute Output, or disclose findings arising from using AlphaFold 3；漏读句（A? 类）：making decisions in domains that affect material or individual rights ；漏读句（D 类）：not (i) resident of a embargoed country, (ii) ordinarily resident in a；漏读句（A? 类）：Google may at any time suspend or terminate your right to use and, as ；漏读句（A 类）：You may only use AlphaFold 3 model parameters if received directly fro" }, "missed_restrictions": [ { "excerpt": "You must not share Output with any commercial organization.", "snapshot_id": "s15", "class": "A", "kind": "restriction" }, { "excerpt": "Distribute Output, or disclose findings arising from using AlphaFold 3 without citing our paper:", "snapshot_id": "s05", "class": "D", "kind": "noise" }, { "excerpt": "making decisions in domains that affect material or individual rights or", "snapshot_id": "s15", "class": "A?", "kind": "unclear" }, { "excerpt": "not (i) resident of a embargoed country, (ii) ordinarily resident in a US", "snapshot_id": "s08", "class": "D", "kind": "noise" }, { "excerpt": "Google may at any time suspend or terminate your right to use and, as applicable", "snapshot_id": "s08", "class": "A?", "kind": "unclear" }, { "excerpt": "You may only use AlphaFold 3 model parameters if received directly from Google.", "snapshot_id": "s02", "class": "A", "kind": "restriction" } ] } }, "access_recipe": { "verified_on": "2026-11-14", "platform": "github", "repo_id": "google-deepmind/alphafold3", "source": { "api_url": "https://api.github.com/repos/google-deepmind/alphafold3", "fetched_on": "2026-11-14" }, "access_state": "application-required", "files": [], "files_truncated": false, "commands": [ { "command": "git clone https://github.com/google-deepmind/alphafold3", "evidence": { "excerpt": "git clone https://github.com/YOUR-USERNAME/YOUR-REPOSITORY", "doc": "github-clone", "url": "https://docs.github.com/en/repositories/creating-and-managing-repositories/cloning-a-repository", "fetched_on": "2026-11-14" } } ], "registration": { "required": true, "url": "https://forms.gle/svvpY4u2jsHEwWYS6", "steps": [ "The model parameters are not downloadable freely: a request form has to be completed and access is granted by the owner (we give no way around this)" ], "evidence": { "excerpt": "To request access to the AlphaFold 3 model parameters, please complete", "snapshot_id": "s14", "url": "https://github.com/google-deepmind/alphafold3/commit/dd1a7badb6" } }, "version": { "commit": null, "last_modified": "2026-11-13T15:55:23Z", "default_branch": "main" }, "latest_release": { "version": "v3.0.5", "date": "2026-10-30T15:31:31Z" }, "license_note": { "license_status": "restricted", "text": "Use according to the terms of the official source; this recipe does not say the data can be used freely." }, "gaps": [ "the platform publishes no checksum for these files; none was computed (we do not download data files)" ] } }, { "slug": "boltz", "name": "Boltz (Boltz-1 and Boltz-2)", "kind": "co-folding", "publisher": "Boltz", "tier": "verified", "license_code": { "status": "permissive", "spdx": "MIT", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "repo-license", "excerpt": "MIT License", "url": "https://raw.githubusercontent.com/jwohlwend/boltz/main/LICENSE", "snapshot_id": "s04", "char_start": 0, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" }, { "location": "repo-license", "excerpt": "MIT License", "url": "https://raw.githubusercontent.com/jwohlwend/boltz/main/LICENSE", "snapshot_id": "s10", "char_start": 0, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "license_weights": { "status": "permissive", "spdx": "MIT", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "dataset-page", "excerpt": "License: mit", "url": "https://huggingface.co/boltz-community/boltz-2", "snapshot_id": "s01", "char_start": 434, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" }, { "location": "dataset-page", "excerpt": "License: mit", "url": "https://huggingface.co/boltz-community/boltz-2", "snapshot_id": "s11", "char_start": 547, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "last_verified": "2026-11-14", "schema_version": "2.0", "record_type": "model", "official_urls": { "repository_or_release": "https://github.com/jwohlwend/boltz", "weights_page": "https://huggingface.co/boltz-community/boltz-2", "paper": null, "license_file": "https://raw.githubusercontent.com/jwohlwend/boltz/main/LICENSE" }, "version_history": [], "output_restrictions": [], "training_data": [], "benchmark_overlap": [], "access": [ { "method": "open-weights-download", "url": "https://huggingface.co/boltz-community/boltz-2", "terms_differ_from_weights_license": false, "own_terms": null } ], "conflicts": [], "known_defects": [], "fit": { "suited_for": [ { "text": { "en": "The README describes Boltz as a family of models for biomolecular interaction prediction.", "zh": "README 将 Boltz 描述为一族用于生物分子相互作用预测的模型。" }, "evidence": [ { "location": "dataset-page", "excerpt": "Boltz is a family of models for biomolecular interaction prediction.", "url": "https://github.com/jwohlwend/boltz", "snapshot_id": "s09", "char_start": 3550, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Boltz repository README", "authors_or_org": "Boltz (jwohlwend/boltz)", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/jwohlwend/boltz" } } ] } ], "not_suited_for": [] }, "gaps": [ { "field": "version_history", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "output_restrictions", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "training_data", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "benchmark_overlap", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "known_defects", "note": "未找到有出处的说明 / no sourced statement found" } ], "completeness": { "core_complete": true }, "changelog": [ { "date": "2026-11-14", "change": { "en": "Added by the automated pipeline from stored page snapshots; every statement carries a verbatim quote.", "zh": "由自动流水线根据已存页面快照添加，每条陈述带逐字引文。" }, "by": "pipeline" } ], "access_recipe": { "verified_on": "2026-11-14", "platform": "huggingface", "repo_id": "boltz-community/boltz-2", "source": { "api_url": "https://huggingface.co/api/models/boltz-community/boltz-2", "fetched_on": "2026-11-14" }, "access_state": "open", "files": [ { "name": ".gitattributes", "url": "https://huggingface.co/boltz-community/boltz-2/resolve/6fdef46d763fee7fbb83ca5501ccceff43b85607/.gitattributes", "size_bytes": 1519, "checksum": null, "format": "gitattributes" }, { "name": "README.md", "url": "https://huggingface.co/boltz-community/boltz-2/resolve/6fdef46d763fee7fbb83ca5501ccceff43b85607/README.md", "size_bytes": 24, "checksum": null, "format": "md" }, { "name": "boltz2_aff.ckpt", "url": "https://huggingface.co/boltz-community/boltz-2/resolve/6fdef46d763fee7fbb83ca5501ccceff43b85607/boltz2_aff.ckpt", "size_bytes": 2062139170, "checksum": { "algorithm": "sha256", "value": "dcc5cd3722b1c9eaa34267e4ae32f55cbbf1963f4c19319381ccfa30fdd2ca9e" }, "format": "ckpt" }, { "name": "boltz2_conf.ckpt", "url": "https://huggingface.co/boltz-community/boltz-2/resolve/6fdef46d763fee7fbb83ca5501ccceff43b85607/boltz2_conf.ckpt", "size_bytes": 2286561469, "checksum": { "algorithm": "sha256", "value": "090e82ac8c92f5e943fa1b39e7410a44027bea7243c0bbb3caa67a77fc1428e1" }, "format": "ckpt" }, { "name": "mols.tar", "url": "https://huggingface.co/boltz-community/boltz-2/resolve/6fdef46d763fee7fbb83ca5501ccceff43b85607/mols.tar", "size_bytes": 1855662080, "checksum": { "algorithm": "sha256", "value": "39e076d96dbec6b4e86982bbda16f3a53a2a60c9bdc17828d88f6f9a0c7d1fd7" }, "format": "tar" } ], "files_truncated": false, "commands": [ { "command": "hf download boltz-community/boltz-2", "evidence": { "excerpt": "You can use the hf download command from the terminal to directly download files from the Hub.", "doc": "hf-download", "url": "https://huggingface.co/docs/huggingface_hub/guides/download", "fetched_on": "2026-11-14" } } ], "registration": null, "version": { "commit": "6fdef46d763fee7fbb83ca5501ccceff43b85607", "last_modified": "2025-06-06T13:18:27.000Z" }, "license_note": { "license_status": "permissive", "text": "Use according to the terms of the official source; this recipe does not say the data can be used freely." }, "gaps": [] } }, { "slug": "esm-c", "name": "ESM C", "kind": "protein-language-model", "publisher": "EvolutionaryScale", "tier": "verified", "license_code": { "status": "permissive", "spdx": "MIT", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "repo-license", "excerpt": "Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction", "url": "https://raw.githubusercontent.com/evolutionaryscale/esm/main/LICENSE.md", "snapshot_id": "s01", "char_start": 64, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" }, { "location": "repo-license", "excerpt": "Copyright 2026 Chan Zuckerberg Biohub, Inc.", "url": "https://raw.githubusercontent.com/Biohub/esm/main/LICENSE.md", "snapshot_id": "s12", "char_start": 19, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "license_weights": { "status": "unverifiable", "spdx": "NOASSERTION", "custom_license": false, "commercial_use": "unverified", "derivatives": "unverified", "redistribution": "unverified", "license_conditions": [], "requires_application": false, "evidence": [ { "location": "dataset-page", "excerpt": "License: other", "url": "https://huggingface.co/EvolutionaryScale/esmc-300m-2024-12", "snapshot_id": "s08", "char_start": 515, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" }, { "location": "dataset-page", "excerpt": "\"license:mit\",\"license:other\"", "url": "https://huggingface.co/biohub/ESMC-300M", "snapshot_id": "s11", "char_start": 1343, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "last_verified": "2026-11-14", "schema_version": "2.0", "record_type": "model", "official_urls": { "repository_or_release": "https://github.com/evolutionaryscale/esm", "weights_page": "https://huggingface.co/EvolutionaryScale/esmc-300m-2024-12", "paper": null, "license_file": "https://raw.githubusercontent.com/evolutionaryscale/esm/main/LICENSE.md" }, "version_history": [ { "version": "unspecified", "released_on": "2025-01-16", "code_license": null, "weights_license": null, "changed": "other", "note": null, "evidence": [ { "location": "dataset-page", "excerpt": "ESM C 600M (incl weights)", "url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md", "snapshot_id": "s13", "char_start": 1063, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Biohub/esm LICENSE.md at commit 1b3e08025b", "authors_or_org": "EvolutionaryScale", "date": "2025-01-16", "date_as_written": "2025-01-16", "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md" } } ] }, { "version": "unspecified", "released_on": "2025-01-16", "code_license": null, "weights_license": null, "changed": "other", "note": null, "evidence": [ { "location": "dataset-page", "excerpt": "Just remember to maintain the same license terms and release under the ESM name.", "url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md", "snapshot_id": "s13", "char_start": 2382, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Biohub/esm LICENSE.md at commit 1b3e08025b", "authors_or_org": "EvolutionaryScale", "date": "2025-01-16", "date_as_written": "2025-01-16", "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md" } } ] }, { "version": "304", "released_on": "2026-05-27", "code_license": null, "weights_license": null, "changed": "other", "note": null, "evidence": [ { "location": "dataset-page", "excerpt": "Update esm repo for ESMC paper ( #304 )", "url": "https://github.com/Biohub/esm/commit/c94ed8d763", "snapshot_id": "s14", "char_start": 2547, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Biohub/esm commit c94ed8d", "authors_or_org": "Chan Zuckerberg Biohub (Biohub/esm)", "date": "2026-05-27", "date_as_written": "2026-05-27", "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/Biohub/esm/commit/c94ed8d763" } } ] } ], "output_restrictions": [ { "kind": "other", "applies_to": "all", "statement": { "text": { "en": "The model card's out-of-scope section prohibits any use barred by the Acceptable Use Policy.", "zh": "模型卡“不适用范围”禁止可接受使用政策所禁止的任何用途。" }, "evidence": [ { "location": "dataset-page", "excerpt": "Any use that is prohibited by the Acceptable Use Policy", "url": "https://huggingface.co/biohub/ESMC-300M", "snapshot_id": "s11", "char_start": 18658, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "ESMC model card — Out-of-Scope or Unauthorized Use Cases", "authors_or_org": "Chan Zuckerberg Biohub (Biohub/esm)", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://huggingface.co/biohub/ESMC-300M" } } ] } } ], "training_data": [ { "resource": "UniRef, MGnify, and the Joint Genome Institute (JGI)", "slug": "uniprot", "role": "pretraining", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "ESM C is trained on protein sequences from UniRef, MGnify, and the Joint Genome Institute (JGI).", "zh": "ESM C 使用来自 UniRef、MGnify 和 Joint Genome Institute (JGI) 的蛋白质序列进行训练。" }, "evidence": [ { "location": "dataset-page", "excerpt": "ESM C is trained on protein sequences from UniRef, MGnify, and the Joint Genome Institute (JGI).", "url": "https://www.evolutionaryscale.ai/blog/esm-cambrian", "snapshot_id": "s02", "char_start": 7326, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } }, { "resource": "UniRef", "slug": "uniprot", "role": "other", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "ESMC was trained on UniRef, MGnify and JGI sequences clustered at 70% identity (83M, 372M and 2B clusters respectively).", "zh": "ESMC 训练数据来自 UniRef、MGnify 与 JGI，按 70% 序列一致性聚类（分别 8300 万、3.72 亿、20 亿簇）。" }, "evidence": [ { "location": "dataset-page", "excerpt": "ESMC was trained on protein sequences from UniRef, MGnify, and the Joint Genome Institute (JGI).", "url": "https://huggingface.co/biohub/ESMC-300M", "snapshot_id": "s11", "char_start": 9479, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "ESMC model card — Training Data", "authors_or_org": "Chan Zuckerberg Biohub (Biohub/esm)", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://huggingface.co/biohub/ESMC-300M" } } ] } } ], "benchmark_overlap": [], "access": [ { "method": "open-weights-download", "url": "https://huggingface.co/EvolutionaryScale/esmc-300m-2024-12", "terms_differ_from_weights_license": false, "own_terms": null }, { "method": "hosted-api", "url": "https://biohub.ai", "terms_differ_from_weights_license": false, "own_terms": null }, { "method": "third-party-hosted", "url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md", "terms_differ_from_weights_license": false, "own_terms": { "text": { "en": "In the 2025 licence, a free non-commercial inference API was offered via Forge (EvolutionaryScale).", "zh": "2025 版许可中，通过 Forge 提供免费的非商业推理 API。" }, "evidence": [ { "location": "dataset-page", "excerpt": "Free non-commercial inference API", "url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md", "snapshot_id": "s13", "char_start": 2468, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Biohub/esm LICENSE.md at commit 1b3e08025b", "authors_or_org": "EvolutionaryScale", "date": "2025-01-16", "date_as_written": "2025-01-16", "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md" } } ] } } ], "conflicts": [], "hardware_notes": [ { "text": { "en": "The model card only recommends a GPU-enabled PyTorch environment, with transformer_engine and xformers optional.", "zh": "模型卡仅推荐带 GPU 的 PyTorch 环境，transformer_engine、xformers 为可选。" }, "evidence": [ { "location": "dataset-page", "excerpt": "PyTorch environment with GPU support recommended.", "url": "https://huggingface.co/biohub/ESMC-300M", "snapshot_id": "s11", "char_start": 9333, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "ESMC model card — System Requirements", "authors_or_org": "Chan Zuckerberg Biohub (Biohub/esm)", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://huggingface.co/biohub/ESMC-300M" } } ] } ], "known_defects": [ { "kind": "bias", "statement": { "text": { "en": "The model card lists dataset bias: over- or under-representation of taxa and protein families in public databases can bias outputs.", "zh": "模型卡列出数据集偏差：公共数据库中物种、蛋白家族的过多或过少代表会使输出产生偏差。" }, "evidence": [ { "location": "dataset-page", "excerpt": "Over- or under-representation of taxa, protein families, or ecological niches in public sequence and structure databases influences generalization and can bias outputs.", "url": "https://huggingface.co/biohub/ESMC-300M", "snapshot_id": "s11", "char_start": 18119, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "ESMC model card — Potential Biases", "authors_or_org": "Chan Zuckerberg Biohub (Biohub/esm)", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://huggingface.co/biohub/ESMC-300M" } } ] }, "source_kind": "official-note", "scope": null, "attribution": { "who": "Chan Zuckerberg Biohub (Biohub/esm)", "when": "date not stated (page read 2026-11-13)", "source_type": "official-page", "title": "ESMC model card — Potential Biases" }, "correction_entrance": "Report an error through the publisher or the authors: https://huggingface.co/biohub/ESMC-300M" } ], "fit": { "suited_for": [ { "text": { "en": "For token or sequence classification, the classifier head is not pretrained and is meant to be fine-tuned for the downstream task.", "zh": "做 token 或序列分类时，分类头未经预训练，需针对下游任务微调。" }, "evidence": [ { "location": "dataset-page", "excerpt": "the classifier head is not pretrained but instead meant to be fine-tuned for your downstream task.", "url": "https://huggingface.co/biohub/ESMC-300M", "snapshot_id": "s11", "char_start": 17017, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "ESMC model card — Usage", "authors_or_org": "Chan Zuckerberg Biohub (Biohub/esm)", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://huggingface.co/biohub/ESMC-300M" } } ] } ], "not_suited_for": [] }, "gaps": [ { "field": "benchmark_overlap", "note": "未找到有出处的说明 / no sourced statement found" } ], "completeness": { "core_complete": true }, "changelog": [ { "date": "2026-11-14", "change": { "en": "Added by the automated pipeline from stored page snapshots; every statement carries a verbatim quote.", "zh": "由自动流水线根据已存页面快照添加，每条陈述带逐字引文。" }, "by": "pipeline" } ] }, { "slug": "esm2", "name": "ESM-2", "kind": "other", "publisher": "Meta AI (FAIR)", "tier": "verified", "license_code": { "status": "permissive", "spdx": "MIT", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "repo-license", "excerpt": "MIT License", "url": "https://raw.githubusercontent.com/facebookresearch/esm/main/LICENSE", "snapshot_id": "s03", "char_start": 0, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "license_weights": { "status": "permissive", "spdx": "MIT", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "dataset-page", "excerpt": "License: mit", "url": "https://huggingface.co/facebook/esm2_t33_650M_UR50D", "snapshot_id": "s01", "char_start": 2992, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "last_verified": "2026-11-14", "schema_version": "2.0", "record_type": "model", "official_urls": { "repository_or_release": "https://github.com/facebookresearch/esm", "weights_page": "https://huggingface.co/facebook/esm2_t33_650M_UR50D", "paper": null, "license_file": "https://raw.githubusercontent.com/facebookresearch/esm/main/LICENSE" }, "version_history": [ { "version": "ESM-2", "released_on": null, "code_license": "MIT", "weights_license": "MIT", "changed": "first-release", "note": null, "evidence": [ { "location": "readme", "excerpt": "- August 2022: ESM-2 - new SOTA Language Models released (see [Lin et al. 2022](https://www.science.org/doi/abs/10.1126/science.ade2574))", "url": "https://raw.githubusercontent.com/facebookresearch/esm/HEAD/README.md", "snapshot_id": "s04", "char_start": 5888, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } ], "output_restrictions": [], "training_data": [ { "resource": "UR50 (sample UR90)", "slug": null, "role": "pretraining", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "The repository page lists ESM-2 checkpoints esm2_t36_3B_UR50D and esm2_t48_15B_UR50D together with the entry UR50 (sample UR90).", "zh": "仓库页面将 ESM-2 检查点 esm2_t36_3B_UR50D 与 esm2_t48_15B_UR50D 同条目 UR50 (sample UR90) 列在一起。" }, "evidence": [ { "location": "readme", "excerpt": "ESM-2 esm2_t36_3B_UR50D() esm2_t48_15B_UR50D() UR50 (sample UR90)", "url": "https://github.com/facebookresearch/esm", "snapshot_id": "s06", "char_start": 8695, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } } ], "benchmark_overlap": [], "access": [ { "method": "open-weights-download", "url": "https://huggingface.co/facebook/esm2_t33_650M_UR50D", "terms_differ_from_weights_license": false, "own_terms": null } ], "conflicts": [], "gaps": [ { "field": "output_restrictions", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "benchmark_overlap", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "known_defects", "note": "未找到有出处的说明 / no sourced statement found" } ], "completeness": { "core_complete": true }, "changelog": [ { "date": "2026-11-14", "change": { "en": "Added by the automated pipeline from stored page snapshots; every statement carries a verbatim quote.", "zh": "由自动流水线根据已存页面快照添加，每条陈述带逐字引文。" }, "by": "pipeline" } ], "access_recipe": { "verified_on": "2026-11-14", "latest_release": null, "files": [], "api": [], "registration": null, "bulk_download": [], "gaps": [ "no direct file address was found on the official pages we hold", "no latest release/version stated on the pages we hold", "the official pages we hold publish no checksum for the files; none was computed (we do not download data files)" ], "license_note": { "license_status": "permissive", "text": "Use according to the terms of the official source; this recipe does not say the data can be used freely." } } }, { "slug": "esm3-open", "name": "ESM3 open (esm3-sm-open-v1)", "kind": "protein-language-model", "publisher": "EvolutionaryScale", "tier": "verified", "license_code": { "status": "permissive", "spdx": "MIT", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "repo-license", "excerpt": "**License (MIT)**", "url": "https://raw.githubusercontent.com/evolutionaryscale/esm/main/LICENSE.md", "snapshot_id": "s03", "char_start": 0, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" }, { "location": "repo-license", "excerpt": "Copyright 2026 Chan Zuckerberg Biohub, Inc.", "url": "https://raw.githubusercontent.com/Biohub/esm/main/LICENSE.md", "snapshot_id": "s06", "char_start": 19, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "license_weights": { "status": "permissive", "spdx": "MIT", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "readme", "excerpt": "This repository is under a MIT [license](https://github.com/evolutionaryscale/esm/blob/main/LICENSE.md).", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/raw/main/README.md", "snapshot_id": "s02", "char_start": 1097, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" }, { "location": "dataset-page", "excerpt": "This repository is under a MIT license", "url": "https://huggingface.co/biohub/esm3-sm-open-v1", "snapshot_id": "s05", "char_start": 2137, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "last_verified": "2026-11-14", "schema_version": "2.0", "record_type": "model", "official_urls": { "repository_or_release": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1", "weights_page": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1", "paper": null, "license_file": "https://raw.githubusercontent.com/evolutionaryscale/esm/main/LICENSE.md" }, "version_history": [ { "version": "ESM-3", "released_on": "2025-01-16", "code_license": null, "weights_license": null, "changed": "other", "note": null, "evidence": [ { "location": "dataset-page", "excerpt": "ESM-3 Open Model", "url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md", "snapshot_id": "s07", "char_start": 887, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Biohub/esm LICENSE.md at commit 1b3e08025b", "authors_or_org": "EvolutionaryScale", "date": "2025-01-16", "date_as_written": "2025-01-16", "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md" } } ] }, { "version": "unspecified", "released_on": "2025-01-16", "code_license": null, "weights_license": null, "changed": "other", "note": null, "evidence": [ { "location": "dataset-page", "excerpt": "Just remember to maintain the same license terms and release under the ESM name.", "url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md", "snapshot_id": "s07", "char_start": 2382, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Biohub/esm LICENSE.md at commit 1b3e08025b", "authors_or_org": "EvolutionaryScale", "date": "2025-01-16", "date_as_written": "2025-01-16", "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md" } } ] }, { "version": "304", "released_on": "2026-05-27", "code_license": null, "weights_license": null, "changed": "other", "note": null, "evidence": [ { "location": "dataset-page", "excerpt": "Update esm repo for ESMC paper ( #304 )", "url": "https://github.com/Biohub/esm/commit/c94ed8d763", "snapshot_id": "s09", "char_start": 2547, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Biohub/esm commit c94ed8d", "authors_or_org": "Chan Zuckerberg Biohub (Biohub/esm)", "date": "2026-05-27", "date_as_written": "2026-05-27", "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/Biohub/esm/commit/c94ed8d763" } } ] } ], "output_restrictions": [ { "kind": "other", "applies_to": "all", "statement": { "text": { "en": "The function decoder was filtered for potentially harmful keywords.", "zh": "功能解码器已过滤潜在有害关键词。" }, "evidence": [ { "location": "dataset-page", "excerpt": "The function decoder has been filtered for potentially harmful keywords.", "url": "https://huggingface.co/biohub/esm3-sm-open-v1", "snapshot_id": "s05", "char_start": 1870, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Model Card for esm3-sm-open-v1", "authors_or_org": "Chan Zuckerberg Biohub (Biohub/esm)", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://huggingface.co/biohub/esm3-sm-open-v1" } } ] } }, { "kind": "other", "applies_to": "all", "statement": { "text": { "en": "The README states the models are available under the MIT licence and asks users to follow Biohub's Acceptable Use Policy.", "zh": "README 称模型以 MIT 许可提供，并要求遵守 Biohub 可接受使用政策。" }, "evidence": [ { "location": "dataset-page", "excerpt": "Please follow our [Acceptable Use Policy](https://biohub.org/acceptable-use-policy/) when using the model.", "url": "https://github.com/Biohub/esm/blob/main/README.md", "snapshot_id": "s08", "char_start": 20053, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Biohub/esm README — Licenses", "authors_or_org": "Chan Zuckerberg Biohub (Biohub/esm)", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/Biohub/esm/blob/main/README.md" } } ] } } ], "training_data": [ { "resource": "2.78 billion natural proteins with synthetic augmentation", "slug": null, "role": "pretraining", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "The model card says the model is trained on 2.78 billion natural proteins, which with synthetic augmentation totaled 3.15 billion protein sequences, 236 million protein structures, 539 million function-annotated proteins and 771 billion tokens.", "zh": "模型卡称该模型基于 27.8 亿个天然蛋白质训练，经合成数据增强后总计 31.5 亿条蛋白质序列、2.36 亿个蛋白质结构、5.39 亿个带功能注释的蛋白质和 7710 亿个 token。" }, "evidence": [ { "location": "readme", "excerpt": "`esm3-sm-open-v1` is trained on 2.78 billion natural proteins. With synthetic data augmentation, this led to 3.15 billion protein sequences, 236 million protein structures, and 539 million proteins with function annotations, totaling 771 billion tokens.", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/raw/main/README.md", "snapshot_id": "s02", "char_start": 90, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } }, { "resource": "training data", "slug": null, "role": "other", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "Trained on 2.78 billion natural proteins, expanded by synthetic augmentation to 3.15 billion sequences, 236 million structures and 539 million function-annotated proteins (771 billion tokens).", "zh": "训练于 27.8 亿条天然蛋白，经合成数据扩增后为 31.5 亿条序列、2.36 亿个结构、5.39 亿条带功能注释蛋白（共 7710 亿 token）。" }, "evidence": [ { "location": "dataset-page", "excerpt": "esm3-sm-open-v1 is trained on 2.78 billion natural proteins.", "url": "https://huggingface.co/biohub/esm3-sm-open-v1", "snapshot_id": "s05", "char_start": 1277, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Model Card for esm3-sm-open-v1", "authors_or_org": "Chan Zuckerberg Biohub (Biohub/esm)", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://huggingface.co/biohub/esm3-sm-open-v1" } } ] } }, { "resource": "training data", "slug": null, "role": "other", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "Virus-related data and some proteins from organisms on the USDA Select Agents and Toxins list were removed from training data.", "zh": "训练数据中移除了病毒相关数据及部分属于 USDA 管制病原与毒素清单的生物蛋白。" }, "evidence": [ { "location": "dataset-page", "excerpt": "data related to viruses has been removed from the training dataset", "url": "https://huggingface.co/biohub/esm3-sm-open-v1", "snapshot_id": "s05", "char_start": 1711, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Model Card for esm3-sm-open-v1", "authors_or_org": "Chan Zuckerberg Biohub (Biohub/esm)", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://huggingface.co/biohub/esm3-sm-open-v1" } } ] } } ], "benchmark_overlap": [], "access": [ { "method": "open-weights-download", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1", "terms_differ_from_weights_license": false, "own_terms": null }, { "method": "third-party-hosted", "url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md", "terms_differ_from_weights_license": false, "own_terms": { "text": { "en": "In the 2025 licence, a free non-commercial inference API was offered via Forge (EvolutionaryScale).", "zh": "2025 版许可中，通过 Forge 提供免费的非商业推理 API。" }, "evidence": [ { "location": "dataset-page", "excerpt": "Free non-commercial inference API", "url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md", "snapshot_id": "s07", "char_start": 2468, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Biohub/esm LICENSE.md at commit 1b3e08025b", "authors_or_org": "EvolutionaryScale", "date": "2025-01-16", "date_as_written": "2025-01-16", "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/Biohub/esm/blob/1b3e08025b/LICENSE.md" } } ] } } ], "conflicts": [], "known_defects": [], "fit": { "suited_for": [ { "text": { "en": "esm3-sm-open-v1 is described as a generative model for designing proteins conditioned on partial sequence, structure and function prompts.", "zh": "esm3-sm-open-v1 被描述为可按序列、结构、功能的部分提示设计蛋白的生成模型。" }, "evidence": [ { "location": "dataset-page", "excerpt": "esm3-sm-open-v1 is a generative model capable of designing proteins conditioned on partial prompts of sequence, structure and function.", "url": "https://huggingface.co/biohub/esm3-sm-open-v1", "snapshot_id": "s05", "char_start": 1530, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Model Card for esm3-sm-open-v1", "authors_or_org": "Chan Zuckerberg Biohub (Biohub/esm)", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://huggingface.co/biohub/esm3-sm-open-v1" } } ] } ], "not_suited_for": [] }, "gaps": [ { "field": "benchmark_overlap", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "known_defects", "note": "未找到有出处的说明 / no sourced statement found" } ], "completeness": { "core_complete": true }, "changelog": [ { "date": "2026-11-14", "change": { "en": "Added by the automated pipeline from stored page snapshots; every statement carries a verbatim quote.", "zh": "由自动流水线根据已存页面快照添加，每条陈述带逐字引文。" }, "by": "pipeline" } ], "access_recipe": { "verified_on": "2026-11-14", "platform": "huggingface", "repo_id": "EvolutionaryScale/esm3-sm-open-v1", "source": { "api_url": "https://huggingface.co/api/models/EvolutionaryScale/esm3-sm-open-v1", "fetched_on": "2026-11-14" }, "access_state": "open", "files": [ { "name": ".gitattributes", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/.gitattributes", "size_bytes": 1519, "checksum": null, "format": "gitattributes" }, { "name": "README.md", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/README.md", "size_bytes": 1202, "checksum": null, "format": "md" }, { "name": "config.json", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/config.json", "size_bytes": 3, "checksum": null, "format": "json" }, { "name": "data/1utn.pdb", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/1utn.pdb", "size_bytes": 568701, "checksum": null, "format": "pdb" }, { "name": "data/ParentChildTreeFile.txt", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/ParentChildTreeFile.txt", "size_bytes": 594786, "checksum": null, "format": "txt" }, { "name": "data/entry_list_safety_29026.list", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/entry_list_safety_29026.list", "size_bytes": 1595245, "checksum": null, "format": "list" }, { "name": "data/esm3_entry.list", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/esm3_entry.list", "size_bytes": 1929090, "checksum": null, "format": "list" }, { "name": "data/hyperplanes_8bit_58641.npz", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/hyperplanes_8bit_58641.npz", "size_bytes": 30026086, "checksum": { "algorithm": "sha256", "value": "25e680def9abb366a31bdea8fe26720828c778dce80689294744afdbe341e13e" }, "format": "npz" }, { "name": "data/hyperplanes_8bit_68103.npz", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/hyperplanes_8bit_68103.npz", "size_bytes": 34870630, "checksum": { "algorithm": "sha256", "value": "886d79883bfc4efb2e160cc327dbdbd25d88bfb3f9774120bb6210cb0ea7efa2" }, "format": "npz" }, { "name": "data/interpro2keywords.csv", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/interpro2keywords.csv", "size_bytes": 7320318, "checksum": null, "format": "csv" }, { "name": "data/interpro_29026_to_keywords_58641.csv", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/interpro_29026_to_keywords_58641.csv", "size_bytes": 10061298, "checksum": null, "format": "csv" }, { "name": "data/keyword_idf_safety_filtered_58641.npy", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/keyword_idf_safety_filtered_58641.npy", "size_bytes": 469256, "checksum": { "algorithm": "sha256", "value": "d83bfb4cda19dcd3bac939c82d369310dbd09b07d30affe7dffc2e4af82200f6" }, "format": "npy" }, { "name": "data/keyword_vocabulary_safety_filtered_58641.txt", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/keyword_vocabulary_safety_filtered_58641.txt", "size_bytes": 787696, "checksum": null, "format": "txt" }, { "name": "data/keywords.txt", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/keywords.txt", "size_bytes": 787695, "checksum": null, "format": "txt" }, { "name": "data/tag_dict_4.json", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/tag_dict_4.json", "size_bytes": 691190, "checksum": null, "format": "json" }, { "name": "data/tag_dict_4_safety_filtered.json", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/tag_dict_4_safety_filtered.json", "size_bytes": 569410, "checksum": null, "format": "json" }, { "name": "data/tfidf_safety_filtered_58641.pkl", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/tfidf_safety_filtered_58641.pkl", "size_bytes": 2022460, "checksum": { "algorithm": "sha256", "value": "36255389d481cffd04f1bcd4a3193360abde685ccafcf03ee26feb50048cfe20" }, "format": "pkl" }, { "name": "data/uniref90_and_mgnify90_residue_annotations_gt_1k_proteins.csv", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/uniref90_and_mgnify90_residue_annotations_gt_1k_proteins.csv", "size_bytes": 109468, "checksum": null, "format": "csv" }, { "name": "data/weights/esm3_function_decoder_v0.pth", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/weights/esm3_function_decoder_v0.pth", "size_bytes": 1299947178, "checksum": { "algorithm": "sha256", "value": "f76d074efcaccfe21365a4fa96f212dadd66798e1e49d809ab7ffbe025d227c9" }, "format": "pth" }, { "name": "data/weights/esm3_sm_open_v1.pth", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/weights/esm3_sm_open_v1.pth", "size_bytes": 2803646068, "checksum": { "algorithm": "sha256", "value": "5ead5a135c658068db6a4f1b933e72d6110992c4668822e1c0e2dcc53e38acd9" }, "format": "pth" }, { "name": "data/weights/esm3_structure_decoder_v0.pth", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/weights/esm3_structure_decoder_v0.pth", "size_bytes": 1237213098, "checksum": { "algorithm": "sha256", "value": "3b726258a44274792b40ce7ea307e10c5da09936368a4ffa2970264d909da65b" }, "format": "pth" }, { "name": "data/weights/esm3_structure_encoder_v0.pth", "url": "https://huggingface.co/EvolutionaryScale/esm3-sm-open-v1/resolve/47f0545b2b6daf26a93439a3cd610f4f7f3d5478/data/weights/esm3_structure_encoder_v0.pth", "size_bytes": 62349276, "checksum": { "algorithm": "sha256", "value": "467acbaee703ba3ccde6e75241a912a316952e5ff071355f85c1d33c68704f40" }, "format": "pth" } ], "files_truncated": false, "commands": [ { "command": "hf download EvolutionaryScale/esm3-sm-open-v1", "evidence": { "excerpt": "You can use the hf download command from the terminal to directly download files from the Hub.", "doc": "hf-download", "url": "https://huggingface.co/docs/huggingface_hub/guides/download", "fetched_on": "2026-11-14" } } ], "registration": null, "version": { "commit": "47f0545b2b6daf26a93439a3cd610f4f7f3d5478", "last_modified": "2026-06-04T17:12:42.000Z" }, "license_note": { "license_status": "permissive", "text": "Use according to the terms of the official source; this recipe does not say the data can be used freely." }, "gaps": [] } }, { "slug": "esmfold", "name": "ESMFold", "kind": "other", "publisher": "Meta AI (FAIR)", "tier": "verified", "license_code": { "status": "permissive", "spdx": "MIT", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "repo-license", "excerpt": "MIT License", "url": "https://raw.githubusercontent.com/facebookresearch/esm/main/LICENSE", "snapshot_id": "s02", "char_start": 0, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "license_weights": { "status": "permissive", "spdx": "MIT", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "readme", "excerpt": "--- license: mit ---", "url": "https://huggingface.co/facebook/esmfold_v1/raw/main/README.md", "snapshot_id": "s03", "char_start": 0, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "last_verified": "2026-11-14", "schema_version": "2.0", "record_type": "model", "official_urls": { "repository_or_release": "https://github.com/facebookresearch/esm", "weights_page": "https://huggingface.co/facebook/esmfold_v1", "paper": null, "license_file": "https://raw.githubusercontent.com/facebookresearch/esm/main/LICENSE" }, "version_history": [], "output_restrictions": [], "training_data": [ { "resource": "PDB + UR50", "slug": "rcsb-pdb", "role": "other", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "The repository README lists PDB + UR50 in the dataset column of the ESMFold row.", "zh": "仓库 README 在 ESMFold 行的数据集列中列出 PDB + UR50。" }, "evidence": [ { "location": "readme", "excerpt": "| ESMFold | `esmfold_v1()` | PDB + UR50 |", "url": "https://raw.githubusercontent.com/facebookresearch/esm/HEAD/README.md", "snapshot_id": "s04", "char_start": 7632, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } } ], "benchmark_overlap": [], "access": [ { "method": "open-weights-download", "url": "https://huggingface.co/facebook/esmfold_v1", "terms_differ_from_weights_license": false, "own_terms": null }, { "method": "hosted-api", "url": "https://esmatlas.com/resources?action=fold", "terms_differ_from_weights_license": false, "own_terms": null } ], "conflicts": [], "gaps": [ { "field": "version_history", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "output_restrictions", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "benchmark_overlap", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "known_defects", "note": "未找到有出处的说明 / no sourced statement found" } ], "completeness": { "core_complete": true }, "changelog": [ { "date": "2026-11-14", "change": { "en": "Added by the automated pipeline from stored page snapshots; every statement carries a verbatim quote.", "zh": "由自动流水线根据已存页面快照添加，每条陈述带逐字引文。" }, "by": "pipeline" } ], "access_recipe": { "verified_on": "2026-11-14", "latest_release": null, "files": [], "api": [], "registration": null, "bulk_download": [], "gaps": [ "no direct file address was found on the official pages we hold", "no latest release/version stated on the pages we hold", "the official pages we hold publish no checksum for the files; none was computed (we do not download data files)" ], "license_note": { "license_status": "permissive", "text": "Use according to the terms of the official source; this recipe does not say the data can be used freely." } } }, { "slug": "evo2", "name": "Evo 2", "kind": "other", "publisher": "Arc Institute", "tier": "verified", "license_code": { "status": "permissive", "spdx": "Apache-2.0", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "repo-license", "excerpt": "Apache License", "url": "https://raw.githubusercontent.com/ArcInstitute/evo2/main/LICENSE", "snapshot_id": "s01", "char_start": 29, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "license_weights": { "status": "permissive", "spdx": "Apache-2.0", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "dataset-page", "excerpt": "License: apache-2.0", "url": "https://huggingface.co/arcinstitute/evo2_7b", "snapshot_id": "s02", "char_start": 466, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "last_verified": "2026-11-14", "schema_version": "2.0", "record_type": "model", "official_urls": { "repository_or_release": "https://github.com/ArcInstitute/evo2", "weights_page": "https://huggingface.co/arcinstitute/evo2_7b", "paper": null, "license_file": "https://raw.githubusercontent.com/ArcInstitute/evo2/main/LICENSE" }, "version_history": [ { "version": "v0.5.0", "released_on": null, "code_license": "Apache-2.0", "weights_license": "Apache-2.0", "changed": "other", "note": null, "evidence": [ { "location": "readme", "excerpt": "**Evo 2 20B released**: 40B-level performance with double the speed, read more [here](https://github.com/ArcInstitute/evo2/releases/tag/v0.5.0).", "url": "https://raw.githubusercontent.com/ArcInstitute/evo2/main/README.md", "snapshot_id": "s03", "char_start": 891, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } ], "output_restrictions": [], "training_data": [ { "resource": "OpenGenome2", "slug": null, "role": "pretraining", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "The README states that Evo 2 was trained autoregressively on OpenGenome2, a dataset containing 8.8 trillion tokens from all domains of life.", "zh": "README 说明 Evo 2 在 OpenGenome2 上以自回归方式训练，该数据集包含来自所有生命域的 8.8 万亿个 token。" }, "evidence": [ { "location": "readme", "excerpt": "Evo 2 was trained autoregressively on [OpenGenome2](https://huggingface.co/datasets/arcinstitute/opengenome2), a dataset containing 8.8 trillion tokens from all domains of life.", "url": "https://raw.githubusercontent.com/ArcInstitute/evo2/main/README.md", "snapshot_id": "s03", "char_start": 443, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } }, { "resource": "Microviridae genomes", "slug": null, "role": "fine-tuning", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "The README lists an evo2_7b_microviridae checkpoint described as a 7B parameter base model fine-tuned on Microviridae genomes.", "zh": "README 列出一个 evo2_7b_microviridae 检查点，描述为在 Microviridae 基因组上微调的 7B 参数基础模型。" }, "evidence": [ { "location": "readme", "excerpt": "| `evo2_7b_microviridae` | 7B parameter base model fine-tuned on Microviridae genomes |", "url": "https://raw.githubusercontent.com/ArcInstitute/evo2/main/README.md", "snapshot_id": "s03", "char_start": 5407, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } } ], "benchmark_overlap": [], "access": [ { "method": "open-weights-download", "url": "https://huggingface.co/arcinstitute", "terms_differ_from_weights_license": false, "own_terms": null }, { "method": "hosted-api", "url": "https://build.nvidia.com/arc/evo2-40b", "terms_differ_from_weights_license": false, "own_terms": null } ], "conflicts": [], "gaps": [ { "field": "output_restrictions", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "benchmark_overlap", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "known_defects", "note": "未找到有出处的说明 / no sourced statement found" } ], "completeness": { "core_complete": true }, "changelog": [ { "date": "2026-11-14", "change": { "en": "Added by the automated pipeline from stored page snapshots; every statement carries a verbatim quote.", "zh": "由自动流水线根据已存页面快照添加，每条陈述带逐字引文。" }, "by": "pipeline" } ], "access_recipe": { "verified_on": "2026-11-14", "latest_release": { "version": "v0.5.0", "date": null, "evidence": { "excerpt": "**Evo 2 20B released**: 40B-level performance with double the speed, read more [here](https://github.com/ArcInstitute/evo2/releases/tag/v0.5.0).", "snapshot_id": "s03", "url": "https://raw.githubusercontent.com/ArcInstitute/evo2/main/README.md" } }, "files": [ { "name": "Evo 2 model checkpoints (HuggingFace)", "url": "https://huggingface.co/arcinstitute", "size_text": null, "checksum": null, "format": null, "version": null, "evidence": { "excerpt": "We provide the following model checkpoints, hosted on [HuggingFace](https://huggingface.co/arcinstitute):", "snapshot_id": "s03", "url": "https://raw.githubusercontent.com/ArcInstitute/evo2/main/README.md" } }, { "name": "OpenGenome2 dataset (HuggingFace)", "url": "https://huggingface.co/datasets/arcinstitute/opengenome2", "size_text": null, "checksum": null, "format": "raw fastas or JSONL files", "version": null, "evidence": { "excerpt": "The OpenGenome2 dataset used for pretraining Evo2 is available on [HuggingFace ](https://huggingface.co/datasets/arcinstitute/opengenome2). Data is available either as raw fastas or as JSONL files which include preprocessing and data augmentation.", "snapshot_id": "s03", "url": "https://raw.githubusercontent.com/ArcInstitute/evo2/main/README.md" } } ], "api": [], "registration": null, "bulk_download": [ { "command_text": "git clone https://github.com/arcinstitute/evo2", "evidence": { "excerpt": "git clone https://github.com/arcinstitute/evo2", "snapshot_id": "s03", "url": "https://raw.githubusercontent.com/ArcInstitute/evo2/main/README.md" } } ], "gaps": [ "the official pages we hold publish no checksum for the files; none was computed (we do not download data files)" ], "license_note": { "license_status": "permissive", "text": "Use according to the terms of the official source; this recipe does not say the data can be used freely." } } }, { "slug": "openfold3", "name": "OpenFold3", "kind": "co-folding", "publisher": "OpenFold consortium", "tier": "verified", "license_code": { "status": "permissive", "spdx": "Apache-2.0", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "repo-license", "excerpt": "Apache License", "url": "https://raw.githubusercontent.com/aqlaboratory/openfold-3/main/LICENSE", "snapshot_id": "s01", "char_start": 34, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "license_weights": { "status": "permissive", "spdx": "Apache-2.0", "custom_license": false, "commercial_use": "yes", "derivatives": "yes", "redistribution": "yes", "license_conditions": [ "notice-required" ], "requires_application": false, "evidence": [ { "location": "dataset-page", "excerpt": "The OpenFold3-preview model is released under Apache 2.0 license.", "url": "https://huggingface.co/OpenFold/OpenFold3", "snapshot_id": "s08", "char_start": 534, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] }, "last_verified": "2026-11-14", "schema_version": "2.0", "record_type": "model", "official_urls": { "repository_or_release": "https://github.com/aqlaboratory/openfold-3", "weights_page": "https://huggingface.co/OpenFold/OpenFold3", "paper": null, "license_file": "https://raw.githubusercontent.com/aqlaboratory/openfold-3/main/LICENSE" }, "version_history": [ { "version": "0.4.2", "released_on": null, "code_license": "Apache-2.0", "weights_license": "Apache-2.0", "changed": "no-license-change", "note": null, "evidence": [ { "location": "readme", "excerpt": "version = {0.4.2}", "url": "https://raw.githubusercontent.com/aqlaboratory/openfold-3/HEAD/README.md", "snapshot_id": "s03", "char_start": 8417, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } ], "output_restrictions": [], "training_data": [ { "resource": "MGnify-based 13M-sequence distillation dataset", "slug": null, "role": "pretraining", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "The README states that the full training data for reproducing OpenFold3-preview includes a reproduction of the MGnify-based 13M-sequence distillation dataset described in the AlphaFold3 paper.", "zh": "README 说明，用于复现 OpenFold3-preview 的完整训练数据包括对 AlphaFold3 论文中所述基于 MGnify 的 1300 万序列蒸馏数据集的复现。" }, "evidence": [ { "location": "readme", "excerpt": "We provide the full training data for reproducing OpenFold3-preview, including our reproduction of the MGnify-based 13M-sequence distillation dataset described in the AlphaFold3 paper.", "url": "https://raw.githubusercontent.com/aqlaboratory/openfold-3/HEAD/README.md", "snapshot_id": "s03", "char_start": 2872, "verified_on": "2026-11-13", "observed": "page", "captured_by": "pipeline" } ] } }, { "resource": "PDB", "slug": "rcsb-pdb", "role": "other", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "OpenBind-0 was trained on PDB data through June 2025 (cutoff 30 June 2025).", "zh": "OpenBind-0 训练数据为截至 2025 年 6 月的 PDB（截止 2025-06-30）。" }, "evidence": [ { "location": "official-change-notice", "excerpt": "OB0 is trained on data from the Protein Data Bank (PDB) through June 2025", "url": "https://openbind.uk/news/blog-openbind-0-advancing-open-molecular-structure-prediction/", "snapshot_id": "s11", "char_start": 1864, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "Blog: OpenBind-0", "authors_or_org": "OpenBind / OpenFold consortium", "date": "2026-08-21", "date_as_written": "2026-08-21", "source_type": "official-change-notice", "open_access": true, "doi_or_url": "https://openbind.uk/news/blog-openbind-0-advancing-open-molecular-structure-prediction/" } } ] } }, { "resource": "MGnify", "slug": null, "role": "other", "cutoff_date": null, "cutoff_stated": false, "statement": { "text": { "en": "Full training data for reproducing OpenFold3-preview are provided, including a reproduction of the MGnify-based 13M-sequence distillation set described in the AF3 paper.", "zh": "提供复现 OpenFold3-preview 的完整训练数据，包括对 AF3 论文所述基于 MGnify 的 1300 万序列蒸馏集的复现。" }, "evidence": [ { "location": "dataset-page", "excerpt": "including our reproduction of the MGnify-based 13M-sequence distillation dataset described in the AlphaFold3 paper.", "url": "https://github.com/aqlaboratory/openfold-3/blob/main/README.md", "snapshot_id": "s09", "char_start": 2941, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "openfold-3 README — Datasets", "authors_or_org": "AlQuraishi Lab / OpenFold consortium", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/aqlaboratory/openfold-3/blob/main/README.md" } } ] } } ], "benchmark_overlap": [], "access": [ { "method": "application-required", "url": "https://huggingface.co/OpenFold/OpenFold3", "terms_differ_from_weights_license": false, "own_terms": null }, { "method": "hosted-api", "url": "https://docs.api.nvidia.com/nim/reference/openfold-openfold3", "terms_differ_from_weights_license": false, "own_terms": null } ], "conflicts": [], "hardware_notes": [ { "text": { "en": "Documentation mentions multi-GPU distributed prediction and a low-memory mode for memory-constrained GPUs; no minimum GPU spec was found.", "zh": "文档提到多 GPU 分布式预测与显存受限时的低显存模式；未见最低 GPU 规格。" }, "evidence": [ { "location": "dataset-page", "excerpt": "Custom settings for memory constrained GPU resources", "url": "https://openfold-3.readthedocs.io/en/latest/", "snapshot_id": "s12", "char_start": 2453, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "OpenFold3-preview documentation", "authors_or_org": "OpenFold Team", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://openfold-3.readthedocs.io/en/latest/" } } ] } ], "known_defects": [], "fit": { "suited_for": [ { "text": { "en": "The maintainers write that OpenBind-0 is advised over OpenFold3-preview2 for everyday usage.", "zh": "维护方写明日常使用建议用 OpenBind-0，而不是 OpenFold3-preview2。" }, "evidence": [ { "location": "dataset-page", "excerpt": "We reccomend using the OpenBind-0 model over OpenFold3-preview2 for everyday usage.", "url": "https://github.com/aqlaboratory/openfold-3/blob/main/README.md", "snapshot_id": "s09", "char_start": 1923, "verified_on": "2026-11-13", "observed": "page", "source_work": { "title": "openfold-3 README", "authors_or_org": "AlQuraishi Lab / OpenFold consortium", "date": null, "date_as_written": null, "source_type": "official-page", "open_access": true, "doi_or_url": "https://github.com/aqlaboratory/openfold-3/blob/main/README.md" } } ] } ], "not_suited_for": [] }, "gaps": [ { "field": "output_restrictions", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "benchmark_overlap", "note": "未找到有出处的说明 / no sourced statement found" }, { "field": "known_defects", "note": "未找到有出处的说明 / no sourced statement found" } ], "completeness": { "core_complete": true }, "changelog": [ { "date": "2026-11-14", "change": { "en": "Added by the automated pipeline from stored page snapshots; every statement carries a verbatim quote.", "zh": "由自动流水线根据已存页面快照添加，每条陈述带逐字引文。" }, "by": "pipeline" } ], "access_recipe": { "verified_on": "2026-11-14", "platform": "huggingface", "repo_id": "OpenFold/OpenFold3", "source": { "api_url": "https://huggingface.co/api/models/OpenFold/OpenFold3", "fetched_on": "2026-11-14" }, "access_state": "gated", "files": [], "files_truncated": false, "commands": [], "registration": { "required": true, "url": "https://huggingface.co/OpenFold/OpenFold3", "steps": [ "The repository is gated: an account and an access request on the repository page are needed (we do not bypass this)" ] }, "version": { "commit": "33ffa390314567dd2d61bb635de52c7810e47056", "last_modified": "2026-05-26T09:32:47.000Z" }, "license_note": { "license_status": "permissive", "text": "Use according to the terms of the official source; this recipe does not say the data can be used freely." }, "gaps": [ "the platform publishes no checksum for these files; none was computed (we do not download data files)" ] } } ] }