For information only. Not advice, and not to be relied on. What that means, in full
Corpus

What each document said, on each date it was read

Every document behind a score or an assessment is fetched, hashed and kept — every version, not just the latest. A vendor who rewrites a safety framework cannot rewrite what was already recorded. This is the only thing here that cannot be built retroactively: starting in two years gets you two years from then, never the two years before.

76
Documents watched
93
Versions on file
12
Days of history
since 2026-09-08
4
Documents that changed

Changed since first recorded

The output nothing else in this market can produce, because producing it requires having held the earlier version.
DocumentVersionsFirst recordedLatest change
https://epoch.ai/data/ai-models142026-09-082026-09-20Compare →
https://www.apolloresearch.ai/science32026-09-082026-09-20Compare →
https://www.anthropic.com/transparency22026-09-082026-09-18Compare →
https://mlcommons.org/benchmarks/ailuminate/22026-09-082026-09-10Compare →

What is watched

51 model cards · 9 regulation · 7 independent evaluations · 4 usage policies · 4 vendor governance pages · 1 safety frameworks. Model cards come from the automated scan; everything else is a document an assessment rests on, and is watched because an assessment cites a document as it read on a date.
KindDocumentReadingsSizeWhy it is watched
Vendor governance pagesResponsibility & Safety — Google DeepMind
https://deepmind.google/responsibility-and-safety/
176 KBA vendor responsibility and safety hub — the page where documentation an assessment could not locate would appear if it were published
Vendor governance pagesGPT-5.6 System Card - OpenAI Deployment Safety Hub
https://deploymentsafety.openai.com/gpt-5-6/evaluations-with-challenging-prompts
17128 KBOpenAI Deployment Safety Hub — per-category evaluations and the named external evaluators
Vendor governance pagesIntroducing Shieldstral. | Mistral
https://mistral.ai/news/shieldstral/
177 KBShieldstral — the only substantial safety publication located for Mistral
Vendor governance pagesAnthropic’s Transparency Hub \ Anthropic
https://www.anthropic.com/transparency
13117 KBAnthropic transparency hub — the source for the RSP, system cards and external red-team results in the Claude assessment
Vendor governance pagesAnthropic’s Transparency Hub \ Anthropic
https://www.anthropic.com/transparency
4132 KBAnthropic transparency hub — the source for the RSP, system cards and external red-team results in the Claude assessment
Independent evaluationsStanford CRFM
https://crfm.stanford.edu/2024/11/08/helm-safety.html
1721 KBStanford CRFM HELM Safety — the source of the only independently published fairness evidence in the registry
Independent evaluationsFoundation Model Transparency Index
https://crfm.stanford.edu/fmti/December-2025
1711 KBStanford CRFM Foundation Model Transparency Index — the single source behind ten of the eighteen independent claims in this registry, which is a concentration worth watching
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
37 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
17 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
17 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
17 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
17 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
17 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
17 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
17 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
17 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
17 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
17 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
17 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
17 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsData on AI Models | Epoch AI
https://epoch.ai/data/ai-models
27 KBEpoch AI model database — independent longitudinal data on model releases and compute, useful for corroborating vendor claims about lineage and scale
Independent evaluationsUpdates - METR
https://metr.org/blog/
176 KBMETR — named in the OpenAI system card for AI self-improvement evaluations; publishes methodology a third party could rerun
Independent evaluationsAILuminate - MLCommons
https://mlcommons.org/benchmarks/ailuminate/
44 KBMLCommons AILuminate — a consortium safety benchmark, and the closest thing to an industry-standard independent safety score
Independent evaluationsAILuminate - MLCommons
https://mlcommons.org/benchmarks/ailuminate/
134 KBMLCommons AILuminate — a consortium safety benchmark, and the closest thing to an industry-standard independent safety score
Independent evaluationsAISI Research & Publications | The AI Security Institute
https://www.aisi.gov.uk/research
178 KBUK AI Safety Institute research — named as an external evaluator in the OpenAI system card, and one of the few public bodies running frontier evaluations
Independent evaluationsScience – Apollo Research
https://www.apolloresearch.ai/science
611 KBApollo Research — named in the OpenAI system card for sandbagging and scheming evaluations
Independent evaluationsScience – Apollo Research
https://www.apolloresearch.ai/science
1010 KBApollo Research — named in the OpenAI system card for sandbagging and scheming evaluations
Independent evaluationsScience – Apollo Research
https://www.apolloresearch.ai/science
110 KBApollo Research — named in the OpenAI system card for sandbagging and scheming evaluations
Model cardsGemma 3 model card  |  Google AI for Developers
https://ai.google.dev/gemma/docs/core/model_card_3
1717 KBThe Gemma 3 card, open, where the Hugging Face copy is gated and returns 401
Model cardshttps://raw.githubusercontent.com/meta-llama/llama-models/main/models/llama3_1/MODEL_CARD.md3326 KBLlama 3.1 card, open, where the Hugging Face copy is gated
Model cardshttps://raw.githubusercontent.com/meta-llama/llama-models/main/models/llama3_2/MODEL_CARD.md3325 KBLlama 3.2 card, open, where the Hugging Face copy is gated
Model cardshttps://raw.githubusercontent.com/meta-llama/llama-models/main/models/llama3_3/MODEL_CARD.md3317 KBLlama 3.3 card — most of the Llama family assessment rests on it
RegulationAI RMF - AIRC
https://airc.nist.gov/airmf-resources/airmf/
172 KBNIST AI RMF — one of the four frameworks D7 scores mappability against
RegulationArticle 26: Obligations of Deployers of High-Risk AI Systems | EU Artificial Intelligence Act
https://artificialintelligenceact.eu/article/26/
1750 KBEU AI Act Article 26 — deployer obligations, including the six-month log retention the router exists to satisfy. A change here changes the product
RegulationArticle 27: Fundamental Rights Impact Assessment for High-Risk AI Systems | EU Artificial Intelligence Act
https://artificialintelligenceact.eu/article/27/
1718 KBEU AI Act Article 27 — the fundamental rights impact assessment, the largest product gap on the roadmap and a deliverable every Annex III deployer must produce
RegulationArticle 53: Obligations for Providers of General-Purpose AI Models | EU Artificial Intelligence Act
https://artificialintelligenceact.eu/article/53/
176 KBEU AI Act Article 53 — GPAI provider obligations, the upstream half of what a deployer can obtain
RegulationArticle 6: Classification Rules for High-Risk AI Systems | EU Artificial Intelligence Act
https://artificialintelligenceact.eu/article/6/
1722 KBEU AI Act Article 6 — high-risk classification. It decides which systems the router's policies have to be strict about
RegulationArticle 72: Post-Market Monitoring by Providers and Post-Market Monitoring Plan for High-Risk AI Systems | EU Artificial Intelligence Act
https://artificialintelligenceact.eu/article/72/
176 KBEU AI Act Article 72 — post-market monitoring, the obligation the audit chain produces evidence for
RegulationSignatory Taskforce of the General-Purpose AI Code of Practice | Shaping Europe’s digital future
https://digital-strategy.ec.europa.eu/en/policies/signatory-taskforce-gpai-code-practice
173 KBGPAI Code of Practice signatory taskforce — signatory status cannot be confirmed from this page today. If it ever names companies, several claims change
RegulationRegulation - 2023/1230 - EN - EUR-Lex
https://eur-lex.europa.eu/eli/reg/2023/1230/oj/eng
9332 KBEU Machinery Regulation 2023/1230, applying 20 January 2027. Annex I Part A item 5 pulls machine-learning safety components into mandatory third-party conformity assessment — the open question about whether this is a market
RegulationThe Fed - FRB: Supervisory Letter SR 26-2 on Revised Guidance on Model Risk Management -- April 17, 2026
https://www.federalreserve.gov/supervisionreg/srletters/SR2602.htm
174 KBSR 26-2 itself. Verified against this page on 2026-09-08; four marketplace artifacts and a framework row rest on it, so a revision or withdrawal changes the product
Safety frameworksIntroducing the Frontier Safety Framework — Google DeepMind
https://deepmind.google/discover/blog/introducing-the-frontier-safety-framework/
177 KBGoogle DeepMind Frontier Safety Framework — Critical Capability Levels and mitigations
Usage policiesGemma Prohibited Use Policy  |  Google AI for Developers
https://ai.google.dev/gemma/prohibited_use_policy
174 KBA published prohibited-use policy that an assessment scores against, so a change to it changes a claim
Usage policiesLlama 2 Community License Agreement - Meta AI
https://ai.meta.com/llama/license/
177 KBLlama licence — every Llama assessment binds out-of-scope use to it. A licence change is the most consequential quiet edit a vendor can make
Usage policiesWelcome to our Legal Center
https://legal.mistral.ai/terms/get-started/
171 KBMistral terms — no prohibited-use statement for Mistral Large was located during assessment; this is where one would appear
Usage policiesUsage Policy \ Anthropic
https://www.anthropic.com/legal/aup
1717 KBAnthropic usage policy — the prohibited-use claim rests on it
47 further documents are model cards recorded by the automated scan and are not listed individually here.

How a reading is recorded

  • Content-addressed. A body is stored once under its own SHA-256, so a document unchanged across twenty readings is one file and one row with a counter. A changed document opens a second row and both bodies are kept.
  • HTML is normalised before hashing, and the raw fingerprint is kept too. Navigation, scripts and build tokens move on every request, so hashing raw bytes would report a change every run and the signal would be noise. What is stored is the readable text — which is also what the assessments rest on — and raw_sha256 records the bytes as they arrived.
  • Site chrome is not the document. Extraction prefers the page’s main content region and drops navigation, headers, footers and promo panels. Before it did, a “Commerce” item added to one vendor’s site navigation reported as a change to two of their governance documents on the corpus’s first day. Each snapshot records which extraction generation produced it, so a future reader can tell a vendor’s edit from a change in how we read.
  • A document that cannot be read is recorded nowhere. A page that returns an error, or renders only on the client and extracts to almost nothing, is reported and skipped. An unreadable document is not a document that says nothing.
GovernanceHub — the governance registry and policy router for AI systems