Anthropic's Mythos and Fable Models Get Green Light for Global Release

The U.S. government has lifted export restrictions on Anthropic's Mythos and Fable AI models, ending a month-long ban that effectively cut off public access to what many consider the most advanced AI systems yet released. Starting Wednesday, July 1, the AI lab will restore access to the models, which had been pulled from public availability after a June 12 ruling placed them on a restricted technology list. That rule required special approval for any foreign national to use the models, a condition Anthropic found impossible to enforce at scale.

Commerce Secretary Howard Lutnick announced the change after weeks of negotiations, noting that Anthropic has agreed to proactively detect and address security risks tied to the models, collaborate with the U.S. government on protocols for future releases, and report any malicious activity. Critics, however, pointed out that Anthropic had already pledged to do most of this voluntarily months before the export rule was imposed. Cybersecurity experts saw the original restrictions less as a genuine security measure and more as leverage - a way for the Trump administration to pressure the company after its executives publicly criticized how the government and political opponents might use AI.

The timing of the reversal also reflects growing competitive pressure from Asian AI firms. Companies like those behind the Fugu and Tulonfeng models are approaching Mythos-level capabilities, pushing Washington to ease restrictions to keep American AI competitive globally. Last week, Lutnick approved limited releases of Mythos to select White House-approved customers, mirroring a similar approach taken with OpenAI's latest models. Industry observers remain wary, however, as the administration's stop-and-start AI policy has left companies uncertain about future rules. An executive order in June calling for pre-release model reviews drew sharp criticism from analysts like Dean W. Ball, who recently joined OpenAI's policy team.

Source: TechCrunch