US Government Orders Anthropic to Disable Claude Fable 5 and Mythos 5

The directive, received by Anthropic at 5:21 pm ET on Friday, mandates a global shutdown of the models for all users. This move stems from export control concerns, though Anthropic claims the government is reacting to a potential jailbreak in Fable 5 that allows the model to identify software vulnerabilities. Anthropic maintains that this capability is already present in other publicly accessible systems, such as OpenAI’s GPT-5.5, and is frequently utilized by cybersecurity professionals for defensive purposes.

Anthropic’s frustration is rooted in the government’s broad application of the ban. The company asserts that its independent classifier systems remain effective at blocking harmful outputs, even if a user attempts to bypass standard refusals. "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 firm stated. They warned that such a strict regulatory standard could effectively paralyze future deployments across the entire frontier AI industry.

This development creates a significant hurdle for Anthropic, which has built its brand on being the safety-conscious alternative to its competitors ahead of a planned IPO. Ironically, the company’s own emphasis on the extreme power of its models may have invited this scrutiny. OpenAI CEO Sam Altman previously dismissed Anthropic’s handling of the Mythos model as "fear-based marketing," suggesting that by positioning their technology as uniquely dangerous, Anthropic inadvertently ensured that regulators would take those threats at face value.