Anthropic Takes Fable 5 and Mythos 5 Offline Over US Export Controls

Anthropic announced Friday that it has taken its latest artificial intelligence models, Fable 5 and Mythos 5, offline to comply with a directive from the Trump administration aimed at preventing their use by foreign nationals. The export controls represent the U.S. government's most significant action to date aimed at restricting access to frontier AI systems, coming just 10 days after President Donald Trump signed an executive order establishing a voluntary federal vetting framework for the most advanced AI models before public release. Under that order, developers can have their most powerful models reviewed for national security risks for up to a month prior to deployment.

Fable 5, released widely earlier this week, is a limited version of the more advanced Mythos model, to which Anthropic has already restricted access due to cybersecurity concerns. The company said it received the directive from the U.S. government Friday afternoon and that the order did not specify the underlying national security concerns. In a statement, Anthropic pushed back on the process, arguing that legitimate restrictions on AI deployment should follow a statutory framework that is "transparent, fair, clear, and grounded in technical facts." The company characterized the situation as a "misunderstanding" and said it is working to restore access to the models "as soon as possible." The Commerce Department did not immediately respond to requests for comment.

The episode highlights the growing intersection of AI governance, national security, and export policy, raising practical questions for developers and enterprises that rely on frontier models. Security teams building on top of these systems should audit their dependencies and consider how abrupt takedowns of major AI services could affect production pipelines. For organizations concerned about data exposure when interacting with AI platforms, running a privacy checkup can help identify inadvertent leaks, while a browser fingerprint test reveals how much identifying information your session may already be exposing to hosted AI services. As regulatory frameworks for AI continue to take shape globally, expect further friction between rapid model deployment and government-mandated review periods, particularly for systems like Mythos 5 that have already drawn scrutiny for capabilities such as rapid exploit generation.