Anthropic says it is disabling Claude Fable 5 and Claude Mythos 5 after receiving a U.S. government export-control directive on June 12, 2026.
The distinction matters: this is not a game being shut down. Fable 5 and Mythos 5 are Claude model names. Fable 5 was Anthropic's public Mythos-class model, launched only days earlier as a safer general-access version of the more restricted Mythos 5 system.
Anthropic says the order requires it to suspend access for any foreign national, including foreign-national Anthropic employees. The company says the practical result is that both models must be disabled for all customers while it complies. Access to other Anthropic models is not affected.
What Anthropic says happened
Anthropic says it received the directive at 5:21 p.m. ET on June 12. The company says the government cited national-security authorities but did not provide specific details of the concern.
Anthropic's understanding is that officials were responding to a claimed method for bypassing, or jailbreaking, Fable 5's safeguards. The company says it reviewed a demonstration and found that it identified a small number of previously known, minor vulnerabilities. Anthropic argues that other publicly available models can discover similar issues without the same bypass.
The company is complying, but it is openly disputing the standard behind the order. Its position is that a narrow, non-universal jailbreak should not be enough to recall a commercial model deployed to a broad user base.
Why this is bigger than one model
The shutdown turns Fable 5 from a model-launch story into a policy story.
Until now, AI export-control fights have mostly focused on chips, data centers, and infrastructure. This order moves the line closer to the deployed model itself: a frontier system that customers were already using can be pulled from production after a government directive.
That creates a new kind of operational risk for teams building around frontier AI. A model can be fast, capable, and available on Monday, then unavailable by Friday for reasons outside the customer's control.
What users should assume
For working professionals, the practical lesson is not only "Anthropic lost access to Fable." It is that model access is now a continuity issue.
Teams using AI for code review, support, research, document work, or internal agents should assume that frontier-model availability can change suddenly. That means fallback models, provider routing, and graceful degradation are becoming part of serious AI operations, not just nice-to-have engineering.
The open questions are still important. Anthropic has not published the government's evidence, the full legal basis for the directive, or a timeline for restoring access. The U.S. government has also not laid out a detailed public explanation matching Anthropic's account.
For now, the confirmed facts are narrower but significant: Fable 5 and Mythos 5 were launched, a U.S. directive followed three days later, and Anthropic says it is disabling both models for all customers while keeping its other Claude models online.
Sources: Anthropic statement, Claude Fable 5 and Claude Mythos 5, Claude support note, TechCrunch, and Business Insider.