{
  "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."
    }
  },
  "_links": {
    "text": "在本库继续查找。本条目是 109 条已处理并核实的记录之一（数据集、数据资源与开源模型）。查找其他条目：调用 `https://mcp.mdexa.com/v1/datasets?q=<关键词>`，或连接 MCP 服务 `https://mcp.mdexa.com/mcp`；每分钟 20 次以内无需密钥。每行返回不到 1 KB；空结果即\"数据库未收录\"。核对一份清单：一次批量核对（免费账号）即返回报告。说明：/agents.md。相似条目：ESM-2 (https://mdexa.com/zh/models/esm2/)；ESMFold (https://mdexa.com/zh/models/esmfold/)；AlphaFold 3 (https://mdexa.com/zh/models/alphafold3/)。",
    "records": 109,
    "search": {
      "template": "https://mcp.mdexa.com/v1/datasets?q={words}",
      "example": "https://mcp.mdexa.com/v1/datasets?q=%E7%BB%93%E5%90%88%E4%BA%B2%E5%92%8C%E5%8A%9B"
    },
    "mcp": "https://mcp.mdexa.com/mcp",
    "anonymous_calls_per_minute": 20,
    "guide": "https://mdexa.com/zh/agents.md",
    "faq": "https://mdexa.com/zh/developers/faq/",
    "similar": [
      {
        "name": "ESM-2",
        "url": "https://mdexa.com/zh/models/esm2/",
        "json": "https://mdexa.com/zh/models/esm2.json"
      },
      {
        "name": "ESMFold",
        "url": "https://mdexa.com/zh/models/esmfold/",
        "json": "https://mdexa.com/zh/models/esmfold.json"
      },
      {
        "name": "AlphaFold 3",
        "url": "https://mdexa.com/zh/models/alphafold3/",
        "json": "https://mdexa.com/zh/models/alphafold3.json"
      }
    ]
  }
}