Anthropic's Fable 5 and Mythos 5 Banned for Foreign Nationals by US Directive
Export control targets Anthropic's frontier models citing a jailbreak vulnerability.
In an unprecedented move, the US government has issued an export control directive under national security authorities to suspend all access to Anthropic's Fable 5 and Mythos 5 models by any foreign national, including foreign national employees of Anthropic, whether inside or outside the United States. The directive was received at 5:21pm ET on June 13, 2026, and requires Anthropic to abruptly disable these models for all customers to ensure compliance. Critically, access to all other Anthropic models will not be affected. The government letter did not provide specific details of its national security concern but indicated it believes a method of bypassing (jailbreaking) Fable 5 has been demonstrated. Anthropic reviewed the technique and found it identifies a small number of previously known, minor vulnerabilities that other publicly-available models can discover without a bypass.
Commentary on LessWrong speculates this is linked to the ongoing fight between Anthropic and the Department of War (DOW), especially after last week's executive order and the failure of an SCR designation to force compliance. While officially issued by the Department of Commerce, many see the DOW's hand behind the move. One commenter noted this is a more devastating tactic than revoking visas of foreign national employees. By disabling the models for US government users as well, Anthropic has escalated the standoff, making further government intervention likely. The situation raises profound questions about government control over frontier AI models and the tension between national security and AI development.
- Directive issued June 13, 2026 at 5:21pm ET by Department of Commerce under national security authorities.
- Only foreign nationals blocked from Fable 5 and Mythos 5, including Anthropic's foreign employees.
- Vulnerability cited is a jailbreak method that finds minor, previously known bugs also found by other models.
Why It Matters
Sets a precedent for government export controls on frontier AI models, intensifying AI safety and sovereignty debates.