The United States has removed a licensing hurdle that had limited Anthropic's ability to share its Mythos and Fable models abroad, reopening access to two of the company's most advanced AI systems.
Anthropic said public availability will begin returning on Wednesday, July 1, after weeks of uncertainty around the models' distribution.
The earlier rule, introduced on June 12, placed the systems on an export-restricted list and blocked access for foreign users without special approval. That framework proved difficult to apply at scale, leading Anthropic to pause public access.
According to Commerce Secretary Howard Lutnick, the company has now agreed to strengthen its approach to security by detecting risks early, coordinating with the U.S. government on standards for current and future releases, and reporting malicious activity when it appears.
Mythos was first offered to a limited set of organizations in April, while Fable later reached the public with added safeguards. The latest policy shift also comes as AI developers in Asia introduce models that are moving closer to Mythos-level capability, increasing pressure for American firms to stay competitive.
With access set to resume, Anthropic's models are once again positioned to shape how advanced AI is deployed, tested, and governed across global markets. The next phase may help define a more balanced future for innovation and oversight.