The U.S. Department of Commerce has officially approved the lifting of export restrictions on Anthropic's latest artificial intelligence models — Claude Fable 5 and Mythos 5. This decision, made on June 30, marks the end of nearly three weeks of intense negotiations between the developer and White House administration officials. Global access to these models will be restored starting July 1.

As a reminder, the restrictions were imposed on June 12, just three days after the official launch of the models on June 9. The administration cited national security concerns, specifically fears regarding potential hacking risks and the use of advanced AI for hostile purposes. Anthropic immediately responded by blocking access to both models for all foreign users across all platforms — Claude.ai, API, AWS Bedrock, and partner services. However, in practice, real-time differentiation of users by citizenship proved technically unfeasible, leading to a complete shutdown of the models for all clients until the situation was clarified.

Why These Models Are So Important

Claude Fable 5 is Anthropic's flagship public model, built on the Mythos architecture but with additional safety layers for mass use. It demonstrates outstanding results in tasks involving complex logic, long-term assignments, engineering calculations, and image processing. Mythos 5, in turn, is designed for more sensitive queries: the same architecture but with loosened cybersecurity restrictions. Access to it was only open to verified partners through the Project Glasswing program for solving government and corporate-level tasks.

The pricing of both models ($10 per million input tokens and $50 per million output tokens) makes them competitive in the market. Built-in classifiers automatically redirect risky queries to safer scenarios, which is especially critical for tasks in cybersecurity and biology.

Cryptalist Expert Opinion: The swift resolution of this conflict sends a powerful signal to the entire industry. It demonstrates that even under strict regulation, dialogue between developers and authorities can be effective. Essentially, we are witnessing the formation of a new regulatory framework for advanced AI in the U.S., where speed of response and willingness to compromise become key factors. For the market, this means reduced uncertainty and, likely, accelerated adoption of such technologies on a global scale.