Commerce Department orders Anthropic to cut Fable 5, Mythos 5 access for all foreign nationals

The US Commerce Department has ordered Anthropic to suspend access to its Fable 5 and Mythos 5 AI models for every foreign national, including the company's own foreign-national staff, citing the risk that the systems could be turned to cyberattacks. Anthropic complied within hours.

The order is a government action, separate from the models' launch earlier this year and the backlash that followed. Commerce Secretary Howard Lutnick sent the directive to Anthropic chief executive Dario Amodei on June 12, 2026, placing the models under export controls for any location outside the United States and for all foreign persons within it. It followed an earlier, unsuccessful push by the Trump administration to get Anthropic to pause releasing the models at all.

The stated concern is technical. Anthropic said on its news page that it received the order at 5:21pm ET and that the letter did not spell out the specific national security concern; the company's understanding is that the government believes it became aware of a method of bypassing, or "jailbreaking," Fable 5. Anthropic says it reviewed a demonstration of the technique being used to identify a small number of previously known, minor vulnerabilities, and that other publicly available models can surface the same flaws without any bypass. Anthropic disabled both models for all customers in response; access to every other Anthropic model was unaffected, the company said.

The scope is unusually broad for an export-control measure. Rather than restricting a specific country or buyer, the directive reaches all non-US persons, a category that includes paying customers abroad and Anthropic employees who are not US citizens. Anthropic, which describes itself as an AI safety company, said it disagrees with the action and believes the finding of a narrow jailbreak should not be cause for recalling a model already deployed to hundreds of millions of people, and that it is working to restore access.

The underlying claim — that the bypass meaningfully raises the odds of real-world cyberattacks — has not been independently verified, and the government has not published technical detail. What is settled is the action itself: a Cabinet secretary directed a leading US AI developer to pull two frontier models from a whole class of users overnight, a precedent likely to shape how Washington polices model access.