At 5:21 p.m. ET on Friday, June 12, Anthropic received an export-control directive from the U.S. Commerce Department ordering Claude Fable 5 and Claude Mythos 5 cut off from any foreign national. By that night the company had disabled both models worldwide. They have been offline for a week. Fable 5, generally available since June 9, lived for three days.

The directive itself, signed by Secretary Howard Lutnick according to Wall Street Journal reporting cited by TIME, hasn't been made public. Anthropic says it was given only verbal evidence of a "potential narrow, non-universal jailbreak" involving prompting the model to read a codebase and fix software flaws. The Journal later identified the underlying paper as the work of Amazon security researchers. Katie Moussouris of Luta Security, who was shown the paper by Anthropic, said it "should never have triggered an export control" and that "the behavior described in the paper cannot meaningfully be fixed, and any attempt would only weaken the model for defense." Anthropic adds, pointedly, that the same prompt works on OpenAI's GPT-5.5.

The company's public posture isn't contrition. "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," Anthropic wrote. "If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers."

Fable 5 is the first publicly deployed Mythos-class model, with a safeguard stack that routes high-risk cybersecurity and biology queries to the older Claude Opus 4.8 rather than Mythos itself. Opus 4.8 and earlier models remain accessible. The capability frontier is, for now, the part that's gone dark.

The political subtext is hard to miss. In February, President Trump ordered federal agencies to stop using Anthropic models after the company refused contract terms permitting use "for any lawful purpose." In early March, the Pentagon designated Anthropic a "supply chain risk," a label Anthropic is challenging in federal court. Axios, as cited by TechCrunch, characterized the export directive as driven by "personality differences" rather than a technical issue. Anton Leicht of the Carnegie Endowment told TIME the directive functions as an "imprecise instrument."

Peter Girnus, a cybersecurity consultant, offered the cleanest read of the broader dynamic to Fortune: "If you describe your product as a munition in every press release, eventually a government takes you at your word." Anthropic's safety-pilled brand built the regulatory vocabulary that's now being used to shelve its flagship. The lesson is that an executive-branch letter, never tested in court, can take a frontier model off the global internet by Friday night.

Commerce pulls Claude Fable 5 and Mythos 5 worldwide over a disputed narrow jailbreak

Sources