Anthropic pulled its Fable 5 and Mythos 5 artificial intelligence models for all customers on June 12 after the U.S. government issued an export control directive suspending foreign nationals’ access to the models, citing national security authorities.
The directive bars any foreign national — inside or outside the United States, including Anthropic’s own foreign national employees — from accessing the two Mythos-class models. To comply, the company said it had to abruptly switch off both models for every user. Access to all other Anthropic models, including Claude for Healthcare, is unaffected.
Anthropic said it received the order at 5:21 p.m. ET on June 12 without specific details of the government’s concern. The company’s understanding is that the government identified a method of “jailbreaking” Fable 5. None of the company’s testers have found a universal jailbreak method, but the company said it’s likely universal jailbreaks will be identified in the future. Anthropic said it reviewed a demonstration of the technique and found it surfaced only a small number of previously known, minor vulnerabilities that other publicly available models, including OpenAI’s GPT-5.5, can also discover. As a result, it’s taking a “defense in depth” strategy with Fable 5.
“We aimed to make jailbreaks either narrow (in the case of non-universal jailbreaks) or very expensive to produce (in the case of universal jailbreaks), and to combine this with thorough monitoring to quickly detect and shut down any successful attacks,” the company said in its statement. “This is also why Anthropic has required 30-day retention of customer data with Fable—a policy change that carries real costs for us with customers, but that allows us to research and mitigate jailbreaks.”
The company said it is complying with the legal directive but disagrees with the action.
“We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people,” the company said in its statement. “If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.”
The recall carries weight for healthcare and life sciences customers. Anthropic had planned a trusted-access program offering Fable 5, with safeguards removed, to select life sciences researchers, and recently said the model cut drug design time tenfold.
The episode also sharpens Anthropic’s move for more coordination and safeguards around advancing AI. The company has pushed for a verifiable pause on frontier AI, arguing that model capabilities are advancing faster than safeguards can keep pace.
