US clears limited use of Anthropic Mythos 5

On June 26, 2026, the U.S. Commerce Department notified Anthropic that it may restore access to its Claude Mythos 5 artificial intelligence model for certain trusted partners. In a letter sent that day to Anthropic’s chief compute officer, Commerce Secretary Howard Lutnick wrote that Anthropic had worked with the U.S. government to address risks associated with its covered models. The letter said these efforts had yielded “significant progress,” clearing the way for a limited redeployment of Mythos 5 rather than a full public release.

The authorization follows earlier government directives that led Anthropic to disable access to its Mythos 5 and Fable 5 models under export control and national security authorities. To comply, the company had been instructed to suspend access for any foreign national, whether inside or outside the United States, including foreign national employees. The new decision marks a partial reopening for Mythos 5 while maintaining tighter overall controls than before the suspension.

The June 26 letter permits release of Mythos 5 to what the Commerce Department describes as “certain trusted partners.” Media reporting indicates this group includes companies and U.S. federal agencies, but the authorization does not extend to a general public rollout. The decision therefore expands availability beyond a complete shutdown while still restricting access to a defined and vetted set of institutions.

The correspondence was characterized as focused on Mythos 5 and did not explicitly change the status of Anthropic’s other powerful system, Fable 5. Reports indicated that discussions over future access to Fable 5 were still underway. As a result, Anthropic’s most advanced models are moving forward under a differentiated regime, with Mythos 5 receiving targeted clearance and Fable 5 remaining under review.

Beyond the immediate access decision, the Commerce Department’s letter underscores a continuing partnership between Anthropic and U.S. authorities on AI safety and governance. The document notes Anthropic’s commitment to work with the government on protocols and standards that will shape how future models are released. This formalizes ongoing engagement around risk assessment, technical safeguards, and deployment conditions for advanced systems.

A Commerce Department spokesperson said the agency had worked diligently to ensure the United States remains a leader in artificial intelligence while safeguarding security. The combination of limited Mythos 5 access and continuing discussions on Fable 5 reflects an approach that links expanded deployment to demonstrated risk‑mitigation measures. For Anthropic and its partners, the June 26 action provides new room to use Mythos 5 within a structured regulatory framework that is still evolving for the company’s most capable AI models.