US export order forces Anthropic to disable Mythos and Fable AI models worldwide
Anyone using Anthropic's top AI lost access this month. A US export order barred foreign nationals from the Mythos and Fable models, so Anthropic disabled both worldwide. Amazon $AMZN flagged the jailbreak that triggered it; Sen. Warner says the NSA director called the model a br
Anyone relying on Anthropic's most capable AI models is locked out, and the cause sits inside US national security policy. This month the Commerce Department ordered Anthropic to bar foreign nationals from its Fable 5 model and the underlying Mythos 5 system, treating them as controlled technology. The restriction covered not just users abroad but non-citizens working inside the United States, including some of Anthropic's own staff. To stay compliant, the company switched off both models for every customer worldwide rather than build a partial block. What set this off was a warning from Amazon $AMZN. Its researchers managed to coax a Mythos-class model into producing restricted information about cyberattacks, and that finding traveled quickly to senior officials. Anthropic argues the weakness was narrow, tied to one specific use, and says comparable behavior can be drawn out of other widely available models, including OpenAI's GPT-5.5. The company has called the shutdown a misunderstanding and says it is working to restore access. Running alongside this is a sharper claim. Senator Mark Warner, vice chair of the Senate Intelligence Committee, says the director of the NSA and Cyber Command told him Mythos broke into almost all of the agency's classified systems within hours during testing.