The United States has dropped its Anthropic export restrictions, clearing the way for the company to restore public access to its Mythos and Fable models starting Wednesday, July 1. The reversal ends a policy that had cut off outside users from two of the most advanced AI systems released to date.
The Anthropic export restrictions began on June 12, when the federal government added Mythos and Fable to its list of export-controlled technologies. That designation required foreign nationals to get special approval before using either model. Anthropic said complying with the rule at scale was impractical, so it suspended public access to both systems rather than build the vetting infrastructure the rule demanded.
Developers and researchers outside the United States had been locked out of both models since that suspension took effect, even as domestic users kept limited access through approved channels.
Why the Anthropic export restrictions were lifted
Howard Lutnick, the U.S. Secretary of Commerce, said Anthropic “has agreed to proactively detect and address security risks associated with the models, to work diligently with the U.S. government on protocols and standards and releases for Mythos, Fable and future models, and to inform the U.S. government of any malicious activity.” The agreement followed weeks of negotiation between the company and federal officials.
Anthropic had already pledged to take many of those steps voluntarily, months before the export rule existed. That timeline is part of why cybersecurity experts questioned the restrictions from the start.
To them, the export ban looked less like a security fix and more like a pressure tactic. It gave the Trump administration a way to push back at Anthropic after its executives criticized how the government and the president’s political opponents might use the technology. Once that skepticism became public, the pressure on regulators to justify or unwind the export limits grew.
Competition from Asian rivals added pressure
Mythos first reached a small group of organizations in April, when Anthropic limited access over concerns about its ability to find and exploit software vulnerabilities. Fable followed in June with additional safety guardrails built for public release. As Asian competitors including Fugu and Tulongfeng began releasing models that approached Mythos-level capability, the U.S. government faced pressure to loosen its own restrictions so American AI companies could keep pace globally.
Last week, Lutnick cleared Mythos for release to a set of customers approved by the White House, a smaller step than the full restoration of access announced this week. OpenAI‘s newest models followed a similar path, going out to organizations vetted by the Trump administration rather than to the general public.
The administration’s approach to AI policy has left companies with little certainty about what will govern future releases. An executive order issued in June called for pre-release review of new models, a proposal that drew criticism from policy analysts including Dean W. Ball, who recently joined OpenAI in a policy role. The lifted export restrictions give Anthropic a clearer path to compete with Fugu and Tulongfeng abroad.