Evaluation evidence is accumulating faster than governance responds. The Governance Lag Monitor is the SaaS intelligence layer for that gap: eval events, framework changes, and regulatory milestones, each tracked to the control or decision that followed, and how long it took to get there.
Anthropic activated AI Safety Level 3 Deployment and Security Standards for Claude Opus 4 while saying it had not yet determined the model definitively passed the ASL-3 capability threshold. The governance response was precautionary: deploy under ASL-3 because ASL-3 risks could not be ruled out.
What happened and when.
The public artifact behind it: system card, published framework version, Model Report, AI Safety Institute publication, regulatory filing.
The control, release decision, escalation, or nothing yet.
Time from evidence to response.
Closed with control, Open, or Watching.
Every entry carries citations.
Each row is a dated evidence item, the governance response that followed, and the lag still open when no response exists.
The public sources are free. The product is the maintained system that turns them into queryable governance objects, watchlists, alerts, and board-ready evidence.
Governance Lag Monitor tracks the response layer: which eval findings became controls, which commitments changed policy, which regulatory clocks are running, and which open loops still need attention.
Queryable tracker, entity views, source chips, and open loop status.
Material changes, new open loops, closed loops, and analyst notes.
Board-ready PDF and institutional source packet.
First edition: Q3 2026.
The preview builds trust. Subscription access gives teams the full working product: search, filters, alerts, exports, briefings, and integrations.
Selected entries, methodology, and public quarterly notes.
Full tracker for people who need to follow governance movement.
Shared watchlists and board-ready material for serious review workflows.
Custom taxonomies, API access, and Syntony advisory support.
Public evidence only. Every entry cites the artifact it is built on: system cards, published frameworks, Model Reports, AI Safety Institute publications, and regulatory filings. Nothing from client engagements, ever.
Open-source research can tell you what was published. The Monitor tells you whether the publication moved governance.
Risk leads, procurement teams, insurers, and boards that need the current state of frontier evaluation and governance response, not last year's scorecard.