U.S. allows Anthropic to release Mythos 5 to trusted organizations after two-week standoff
Tags AI · Enterprise · Policy
The Trump administration granted Anthropic permission to release Mythos 5 to a select group of vetted US organizations, following a two-week negotiation that began when Anthropic took its Mythos-class models offline on a Friday evening ultimatum. Commerce Secretary Howard Lutnick sent a letter dated June 26 to Anthropic co-founder Tom Brown confirming a revision to license requirements. The public-facing Fable 5 model remains in limbo with no timeline for broader rollout.
Technical significance
The restricted release framework for Mythos 5 establishes a new government gatekeeping model for frontier AI deployments, where access is determined by federal vetting rather than market dynamics. This creates a bifurcated ecosystem: cleared organizations gain access to the most capable models while others remain on older versions, potentially widening the AI capability gap between vetted and non-vetted entities.