Anthropic cuts off Fable 5 and Mythos 5 after government national security order
Government forces Anthropic to block its top models for all foreign nationals worldwide.
On Friday evening, the U.S. government ordered Anthropic to block access to its advanced AI models Fable 5 and Mythos 5 for all foreign nationals, both inside and outside the U.S., due to national security concerns. The directive also applied to Anthropic’s own foreign-national employees. To comply, Anthropic completely shut down access to these models for all customers.
In a statement, Anthropic said it complied but criticized the government for not providing specific details of the national security threat. The company claims any evidence of potential jailbreak was shared verbally, and the vulnerabilities discovered were minor and not unique to Mythos 5—similar issues exist in other models like GPT-5.5. Anthropic also noted it had already worked with U.S. and UK governments on safeguards, including updated data retention policies to track malicious use. The company emphasized it had not received disclosure of any harmful, non-universal jailbreak that would justify the order.
- Government ordered Anthropic to block Fable 5 and Mythos 5 for all foreign nationals, including employees.
- Anthropic complied fully but says no specific national security details were provided; only verbal evidence of 'minor' jailbreaks.
- Company claims similar vulnerabilities exist in models like GPT-5.5, and no non-universal harmful jailbreak was disclosed.
Why It Matters
Sets a precedent for government intervention in AI model access over national security claims, impacting global deployment.