The US government has eased restrictions on Anthropic’s advanced AI model, Claude Mythos 5, allowing it to be accessed by over 100 select US organizations. However, a broader rollout remains restricted.

Despite the partial reinstatement, concerns about security and foreign access persist. The saga raises questions about future AI policy under the Trump administration.

Following a series of incidents, including a South Korean firm’s potential ties to China and separate concerns from Amazon and the NSA over Fable 5's potential jailbreaking, Anthropic faced significant delays and even legal battles with the government.

The company is now working to provision approved providers quickly while continuing discussions with the White House to restore broader access. The incident has also highlighted the need for explicit government approval for AI model releases in the future.