Trump Admin Allows Anthropic's Mythos 5 to 100+ US Organizations After Ban
After two-week ban, Anthropic's strongest cybersecurity model returns for critical infrastructure.
Two weeks after imposing a ban that forced Anthropic to pull its powerful cybersecurity models—Mythos 5 and Fable 5—the Trump administration is now easing restrictions. In a directive seen by Semafor, Commerce Secretary Howard Lutnick authorized Anthropic to redeploy Mythos 5 to more than 100 specific U.S. government agencies and companies, including allowing non-American employees at those organizations to access the model. The same exemption applies to Anthropic’s own non-American employees, who were originally barred under the ban.
The decision follows concerns that security researchers had easily bypassed the models’ guardrails, leading to the initial pull. Anthropic publicly acknowledged the development on X, stating it has been working with the government since June 12 to restore access. While Mythos 5 is now available for critical infrastructure defense, Fable 5 remains restricted. Anthropic continues to seek broader approval for both models, aiming for general availability again.
- Over 100 U.S. government agencies and companies granted access to Mythos 5, including non-American employees.
- Original ban was triggered after security researchers bypassed Mythos 5’s guardrails with relative ease.
- Fable 5, the more protected variant, remains restricted and was not included in the new directive.
Why It Matters
Restores powerful AI cybersecurity tool to critical infrastructure defenders while balancing national security risks and access for non-U.S. personnel.