{
  "dataset": "safety-frameworks",
  "title": "Frontier safety frameworks",
  "index": "Frontier Safety Framework Index",
  "url": "https://aisafetytracker.org/alignment/#frameworks",
  "citation": "AI Safety Tracker, \"Frontier Safety Framework Index\". https://aisafetytracker.org/alignment/#frameworks. Retrieved 23 October 2026.",
  "licences": [
    {
      "source": "safety-frameworks-index",
      "name": "Frontier Safety Framework Index",
      "publisher": "This site",
      "licence": "MIT",
      "licence_url": "https://opensource.org/licenses/MIT",
      "attribution": "AI Safety Tracker, \"Frontier Safety Framework Index\". Licensed MIT."
    }
  ],
  "generated": "2026-10-23T21:53:16+00:00",
  "retrieved": "2026-10-23T00:00:00+00:00",
  "unit": null,
  "notes": "Our own reading of each developer's published frontier safety framework, reviewed 2026-10-23. Self-assessment under self-written frameworks with no common scale and no external audit. Not a ranking.",
  "count": 8,
  "records": [
    {
      "developer": "Anthropic",
      "framework": "Responsible Scaling Policy",
      "scale": "Four capability thresholds, each paired with recommended mitigations; AI Safety Levels now name present safeguards only",
      "levels": 0,
      "domains": [
        "CBRN",
        "loss of control",
        "AI R&D"
      ],
      "external": true,
      "halt": "qualified",
      "since": 2023,
      "url": "https://www-cdn.anthropic.com/files/4zrzovbb/website/0bacdc8440ea96e62a8766d99ebe1d4eea6d5f3a.pdf",
      "note": "Rewritten as version 3 in February 2026; version 3.4 took effect on 8 July 2026. The earlier commitment not to train or deploy without adequate safeguards is gone: it commits to delay development and deployment only while Anthropic leads or its competitors have strong safety measures, and otherwise publishes Risk Reports with a risk-benefit determination by its chief executive and Responsible Scaling Officer. It no longer has a cyber threshold.",
      "halt_meaning": "Commits to a decision process rather than to an outcome.",
      "reviewed": "2026-10-23",
      "source_id": "safety-frameworks-index"
    },
    {
      "developer": "OpenAI",
      "framework": "Preparedness Framework",
      "scale": "Capability thresholds: High and Critical",
      "levels": 2,
      "domains": [
        "CBRN",
        "cyber",
        "AI R&D"
      ],
      "external": true,
      "halt": "yes",
      "since": 2023,
      "url": "https://cdn.openai.com/pdf/18a02b5d-6b67-4cec-ab64-68cdfbddebcd/preparedness-framework-v2.pdf",
      "note": "Version 2, April 2025, is still the published version. OpenAI's Frontier Governance Framework of May 2026 sets out how it meets EU and Californian law and leaves this as its own standard. Three tracked categories at two thresholds: persuasion was dropped in 2025, and long-range autonomy and nuclear risk are research categories without thresholds. At a Critical threshold it commits to halt further development until safeguards are specified.",
      "halt_meaning": "Commits to not deploying, or not continuing, until mitigations are in place.",
      "reviewed": "2026-10-23",
      "source_id": "safety-frameworks-index"
    },
    {
      "developer": "Google DeepMind",
      "framework": "Frontier Safety Framework",
      "scale": "Critical Capability Levels per domain, with lower Tracked Capability Levels for some",
      "levels": 0,
      "domains": [
        "CBRN",
        "cyber",
        "manipulation",
        "AI R&D",
        "loss of control"
      ],
      "external": false,
      "halt": "qualified",
      "since": 2024,
      "url": "https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/strengthening-our-frontier-safety-framework/frontier-safety-framework_3-1.pdf",
      "note": "Version 3.1 took effect on 17 April 2026. Adds Tracked Capability Levels below the critical ones, for CBRN and for a domain that now merges machine-learning R&D with misalignment. Risk acceptance weighs what other publicly available models can do, and for misuse risks applies to external deployment only, not to internal use or further development.",
      "halt_meaning": "Commits to a decision process rather than to an outcome.",
      "reviewed": "2026-10-23",
      "source_id": "safety-frameworks-index"
    },
    {
      "developer": "Meta",
      "framework": "Superintelligence Scaling Framework",
      "scale": "Risk thresholds: moderate or lower, high, critical",
      "levels": 3,
      "domains": [
        "CBRN",
        "cyber",
        "loss of control"
      ],
      "external": false,
      "halt": "yes",
      "since": 2025,
      "url": "https://research.meta.ai/static/meta-superintelligence-scaling-framework",
      "note": "Version 2.1, 2 October 2026, renamed from the Advanced AI Scaling Framework of April 2026, itself renamed from the Frontier AI Framework. Since version 2 the critical tier reads Develop with Mitigations, formerly Stop, so every tier permits development and deployment once mitigations are validated to bring risk down to moderate. Version 2.1 adds a third loss-of-control outcome, containment failure in training and evaluation. Covers any model trained with 1e26 FLOP or more.",
      "halt_meaning": "Commits to not deploying, or not continuing, until mitigations are in place.",
      "reviewed": "2026-10-23",
      "source_id": "safety-frameworks-index"
    },
    {
      "developer": "Microsoft",
      "framework": "Frontier Governance Framework",
      "scale": "Risk levels per capability: low, medium, high, critical",
      "levels": 4,
      "domains": [
        "CBRN",
        "cyber",
        "AI R&D",
        "loss of control",
        "manipulation"
      ],
      "external": true,
      "halt": "yes",
      "since": 2025,
      "url": "https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/msc/documents/presentations/CSR/Frontier-Governance-Framework-Feb-2026.pdf",
      "note": "Revised in February 2026 to align with the EU Code of Practice and the Californian and New York frontier AI laws, adding loss of control and harmful manipulation. Commits to pause development and deployment of a model whose risk cannot be sufficiently mitigated. Covers models in scope of those laws, and Microsoft's substantial fine-tunes of other developers' models.",
      "halt_meaning": "Commits to not deploying, or not continuing, until mitigations are in place.",
      "reviewed": "2026-10-23",
      "source_id": "safety-frameworks-index"
    },
    {
      "developer": "Amazon",
      "framework": "Frontier Model Safety Framework",
      "scale": "Critical capability thresholds",
      "levels": 0,
      "domains": [
        "CBRN",
        "cyber",
        "loss of control",
        "manipulation"
      ],
      "external": true,
      "halt": "yes",
      "since": 2025,
      "url": "https://www.amazon.science/publications/amazons-frontier-model-safety-framework",
      "note": "Revised in September 2026. Names four critical risk domains, adding harmful manipulation, and commits not to deploy a model that meets a threshold until safeguards appropriately mitigate the risks.",
      "halt_meaning": "Commits to not deploying, or not continuing, until mitigations are in place.",
      "reviewed": "2026-10-23",
      "source_id": "safety-frameworks-index"
    },
    {
      "developer": "xAI",
      "framework": "Frontier Artificial Intelligence Framework",
      "scale": "Benchmark thresholds for deployment, in two domains",
      "levels": 0,
      "domains": [
        "CBRN",
        "loss of control"
      ],
      "external": false,
      "halt": "qualified",
      "since": 2025,
      "url": "https://data.x.ai/2025-12-31-xai-frontier-artificial-intelligence-framework.pdf",
      "note": "Version of December 2025, written to comply with California's SB 53. Sets two deployment criteria: an answer rate below 1 in 20 on restricted biology and chemistry queries, and a dishonesty rate below 1 in 2 on the MASK benchmark. Cyber misuse is handled by a refusal policy, with no threshold. States that the benefits of deployment may outweigh a risk a benchmark identifies.",
      "halt_meaning": "Commits to a decision process rather than to an outcome.",
      "reviewed": "2026-10-23",
      "source_id": "safety-frameworks-index"
    },
    {
      "developer": "DeepSeek",
      "framework": null,
      "scale": null,
      "levels": 0,
      "domains": [],
      "external": false,
      "halt": "no-framework",
      "since": null,
      "url": "https://www.gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024",
      "note": "Listed as an absence. DeepSeek releases open-weight models at or near the frontier and has published no safety framework, and is not among the Seoul signatories. The linked document is the commitment list it is absent from.",
      "halt_meaning": "No published framework, so nothing to commit to.",
      "reviewed": "2026-10-23",
      "source_id": "safety-frameworks-index"
    }
  ]
}