The U.S. government has activated export control guidelines that fully prohibit foreign nationals from accessing Anthropic’s latest artificial intelligence models, 'Fable5' and 'Mythos5'. Anthropic blocked access for all customers to comply, but objected, saying the measure stemmed from a misunderstanding.

On June 12 local time, Anthropic said guidance it received at 5:21 p.m. Eastern time cited national security and stipulated a ban on access to Fable5 and Mythos5 for all foreign nationals. It includes foreigners staying in the United States and foreign employees at Anthropic. Other Anthropic models are not affected by the measure.

According to U.S. political media outlet Axios, the guidance was delivered in the form of a letter from U.S. Commerce Secretary Howard Lutnick to Anthropic CEO Dario Amodei (다리오 아모데이). Axios reported that the Commerce Department issued the export control guidance after a company other than Anthropic claimed it had found a method to jailbreak Mythos, which the department judged to be a security threat.

Axios added that the U.S. government had previously tried to halt the release of Anthropic’s models but did not succeed, leading to the latest measure.

Axios reported, citing remarks by a government official, that the access restrictions would remain in place for several weeks.

Anthropic disputed the move, saying it was a misunderstanding. In an official statement, it said, "So far the government has provided only verbal reports without concrete evidence of jailbreaking, and demonstrations we reviewed showed the vulnerability could be reproduced in the same way in other public models, including OpenAI’s GPT-5.5, and is at a level security experts use every day."

It added, "We do not agree that a commercial model distributed to hundreds of millions of people should be recalled solely because a narrow set of potential jailbreaking methods was found," and "If this standard is applied across the industry, new deployment of all frontier models would effectively be halted."

Anthropic also said, "The government should have the authority to block unsafe AI deployments through transparent, fair legal procedures grounded in technical facts," but added, "This measure does not align with those principles." It also viewed the access shutdown as "an extension of recent conflict with the U.S. government."

Previously, the U.S. Department of Defense designated Anthropic as a supply chain risk company and excluded it from government procurement.