US lifts export curbs on Anthropic AI models Fable 5, Mythos 5

The United States government has lifted export controls on AI firm Anthropic's 'Fable 5' and 'Mythos 5' models, signaling a significant shift in its approach to regulating advanced artificial intelligence. The restrictions, imposed earlier in May due to national security concerns, had blocked foreign access to these powerful AI systems.

“

We received notification from the U.S. Department of Commerce that export controls on Claude 'Fable 5' and 'Mythos 5' have been lifted.

Anthropic announced on its official X account that it received notification from the U.S. Department of Commerce to lift the export controls. Access is expected to be restored sequentially starting May 31. The company expressed gratitude to its users and collaborators for their patience and support during the period of restricted access.

The lifting of these restrictions is contingent on Anthropic's commitment to work with the Commerce Department. According to a letter from Commerce Secretary Howard Lutnick, Anthropic pledged to develop protocols and standards for its AI models, ensuring proactive measures against potential security risks. Mythos 5 is a specialized cybersecurity model for businesses, while Fable 5 is a version adapted for general users. Anthropic had previously delayed the release of Mythos 5, deeming it "too powerful."

“

We thank our users for their patience and everyone who worked with us to redeploy the models.

This development occurs as low-cost, high-performance AI models from China rapidly gain traction. Industry observers suggest that concerns raised by major AI companies about these emerging Chinese models may have influenced the U.S. government's decision to ease restrictions. Some investors and tech executives have voiced worries that U.S. regulations could hinder domestic companies' competitiveness in the global technology race. OpenAI, for instance, recently limited initial access to its new model, 'GPT-5.6,' to a select group of trusted users at the government's request, prompting criticism that such practices should not become a government standard.

“

Anthropic will establish protocols and standards to be applied to AI models in the future and will proactively block and respond to security risks related to the models.