Anthropic said on Friday that it disabled access to its Fable 5 and Mythos 5 artificial intelligence models after receiving a U.S. government export control order citing national security authorities. The company said the directive, delivered at 5:21 p.m. ET, instructed it to suspend access to the models for any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. Anthropic said it complied immediately and took the models offline for all customers to ensure it met the order, while confirming that its other AI models were not affected.
The decision came only days after Anthropic unveiled Fable 5 and Mythos 5, describing both as highly capable systems that led industry benchmarks. Fable 5 was presented as a major public release, with new safety measures designed to block answers in certain high-risk areas. Mythos 5 was based on Claude Mythos Preview, a model that drew attention from Wall Street and government officials earlier in the year because of its advanced cybersecurity abilities.
Anthropic said the models were not intended for broad release in the first place. The company had limited access to a small number of firms through Project Glasswing, a cybersecurity initiative. Still, the sudden shutdown affected customers already using the models and disrupted a product rollout that had only recently begun.
In its statement, Anthropic said the government had not explained the specific national security concern behind the order. The company apologized for the interruption and said it supports the government’s ability to stop unsafe AI deployments, but only through a process that is transparent, fair, clear, and based on technical facts. Anthropic said the action it received did not meet those standards.
The move adds to Anthropic’s tensions with the U.S. government. Earlier this year, the company became embroiled in a public dispute with the Department of Defense after negotiations broke down. The Pentagon later labeled Anthropic a supply chain risk, a designation that has traditionally been used for foreign adversaries and requires defense contractors to certify they will not use Anthropic’s Claude models in military-related work.
Anthropic has since sued the Trump administration to challenge that blacklisting, and the case is still pending. The latest export control action underscores the growing scrutiny facing advanced AI systems as U.S. officials weigh their potential benefits against cybersecurity and national security risks.
The company said the government’s order will not affect its other models, but the incident highlights how quickly AI products can become entangled in policy and security disputes as the technology advances and expands into sensitive use cases.
