Anthropic's account indicates a material safety-control and access-governance response, but the effectiveness figures and comparative capability claims remain primarily self-reported.
Limits of the evidence: The source is Anthropic's own account; the reported over-99% blocking rate, model comparisons, and government and CAISI assessments are not independently corroborated in the supplied text.
What to watch
Taken from the first report; the whole story has not been re-assessed yet.
Redeploying Fable 5
AnthropicJul 6, 2026First report
No revisions yet.
No maturity ladder applies to this desk.
Sens.ai aggregates and assesses published reporting. The assessment above is machine generated; the original sources are authoritative.