The U.S. government has ordered Anthropic to immediately suspend access to two of its most powerful AI models — Claude Fable 5 and Mythos 5 — for all users who are not U.S. citizens. Anthropic received the directive and complied the same day, cutting off foreign access while the situation is assessed. The government cited national security concerns, saying it has evidence of a technique that can bypass Fable 5’s built-in safety protections.
Mythos 5, a version of the same underlying model with reduced safety guardrails designed for cybersecurity professionals, is also suspended. Other Claude models — including Claude Sonnet and Haiku — remain available as normal. Anthropic has publicly pushed back on the scope of the order, calling the jailbreak method “relatively simple” and comparable to techniques that already work on other AI models from companies like Google and OpenAI. The company says it is working to restore access and believes the order may be based on a misunderstanding.
How to check if you’re affected
Affected products include Claude Fable 5 and Mythos 5. If you rely on these models through the Anthropic API or Claude.ai and you are not a U.S. citizen, you will find your access has been suspended. Log into your Anthropic account and check which models are available — if Fable 5 or Mythos 5 are greyed out or missing, your access has been restricted. Users of other Claude models are unaffected.