Anthropic disables Fable 5, Mythos 5 after US export order

Anthropic will disable access to its two newest AI models, Fable 5 and Mythos 5, after the US government issued an export-control order citing national security risks. The company said the order would bar foreign individuals and entities from using the models. The exact compliance deadline was not disclosed.

The move marks the first time a major US AI lab has publicly disabled a frontier model in response to direct government restrictions. Previous controls targeted chip exports and cloud-computing access, not the model weights themselves. Anthropic’s decision signals that export enforcement now extends to the software layer of artificial intelligence.

The order comes from the Commerce Department’s Bureau of Industry and Security, which has broad authority to restrict exports of “emerging and foundational technologies.” Fable 5 and Mythos 5 are the most advanced models Anthropic has released. They power a range of applications from coding assistants to enterprise chatbots. The restriction effectively blocks users in countries subject to US arms embargoes, as well as entities deemed a national security risk.

Anthropic did not say whether it will offer a less capable version for international customers. The company relies on API subscriptions and enterprise licensing for revenue. A partial or regional block would narrow its addressable market at a time when competition from OpenAI and Google is intensifying.

The US government has not publicly explained the specific threat posed by the two models. Officials have cited concerns that frontier AI could accelerate weapons development or automate cyberattacks. Anthropic has voluntarily disclosed safety evaluations to the government before each major release. That process did not prevent the export order.

Industry groups have warned that overlapping controls create uncertainty for developers. The order applies to the model weights and the trained parameters, not just the source code. Companies must determine on their own whether a foreign user’s request violates the terms. That shifts compliance costs from the government to the private sector.

Anthropic said it will notify affected customers and provide a timeline for disabling access. The company declined to comment on whether it challenged the order before complying. A spokesperson said Anthropic supports the goal of national security but did not elaborate on the specific disagreement.

The broader impact on the AI industry will depend on how other labs respond. If similar orders follow for other frontier models, access to cutting-edge AI could become fragmented along national lines. For now, the immediate question is whether international users of Fable 5 and Mythos 5 can migrate to older models or alternative providers before the cutoff.