Anthropic announced on June 26 that the U.S. government has approved limited access to Mythos 5, its most advanced AI model, for certain organizations. The partial reversal follows a June 12 export-control directive that suspended both Mythos 5 and the general-purpose Fable 5 over jailbreak concerns.
The government initially blocked the models after identifying a method to circumvent Fable 5's safety guardrails, a process known as jailbreaking. The latest decision allows Mythos 5, described by Anthropic as its strongest cybersecurity model, to be redeployed to a small group of cyber defenders and infrastructure providers.
Anthropic said it is working to restore access for approved providers as quickly as possible. The status of Fable 5 remains unchanged, with no indication of when general access might resume. The company continues discussions with officials to expand Mythos 5's availability and bring back Fable 5.
This limited release signals a cautious approach to balancing national security with commercial AI development. Critics on the right argue the restrictions hamper U.S. competitiveness against China, while security experts warn that premature deployment could expose critical infrastructure to sophisticated attacks.
"We're continuing to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again," Anthropic said in its statement, emphasizing ongoing negotiations.