Outside red-teams of the controls
Reference classes
OpenAI current estimate (holding)
9/10 (0.50)
Anthropic current estimate (inferred)
5/10 (0.50)
Resolution
Source: OpenAI & Anthropic system cards, Preparedness / RSP updates, or independent audits
Includes: Public primary-source confirmation that Outside red-teams of the controls is operating across frontier deployments. Assurance / red-team evaluation on Internal deployments (narrows uncertainty; zero separate blocking credit).
Excludes: Aspirational roadmap commitments, offline toy-model pilots, or unverified third-party claims.