Threshold Watch

Every frontier lab publishes a safety framework: the capability thresholds that are supposed to trigger stronger safeguards. None of them publishes a version history. The thresholds differ between labs, change without notice, and changes are how commitments soften. This page does the versioning for them: each framework is fetched on a cadence, normalized, hashed, and archived. When the content changes, the new snapshot lands here, dated, with a diff where the source is text.

As of September 2026, Threshold Watch tracks six frontier labs (Anthropic, OpenAI, Google DeepMind, xAI, Meta, Amazon). All six publish a safety framework. None publishes a changelog a third party can verify, and one framework (Meta's) rejects automated retrieval outright.

This is commitment one of the Control Stack Compact run in reverse: thresholds should be published in a common format and versioned like a changelog by the labs themselves. Until they are, the changelog lives here.

Subscribe to changes: RSS feed (a new item means a framework changed or entered tracking). Related: the Frontier Incident Timeline and the Compact Clauses for procurement.

For reporters: when a lab edits its safety framework, the change lands here first, as a dated snapshot with a hash and a line-level diff against the previous version. Subscribe to the RSS feed above to see edits the day they are captured, and cite the diff, not our summary. Questions or verification requests: hello@majormatters.co.