Anthropic announced that the US Department of Commerce has removed export restrictions on its AI models, Fable and Mythos, just under three weeks after the company was directed to halt access to its more sophisticated models due to security concerns.
In a post on X late Tuesday, Anthropic confirmed, “We’ll begin restoring access tomorrow.”
The US government's decision to block access to these models was rooted in national security issues, but a letter from US Commerce Secretary Howard Lutnick, which was obtained by Reuters, lifted these restrictions, indicating that no license is now necessary for export.
Lutnick highlighted that Anthropic has committed to actively identify and mitigate security risks tied to the models. The company will collaborate closely with the US government to establish protocols and standards for Mythos, Fable, and future models, as well as report any malicious activities associated with them.
Amid growing awareness of potential threats, the US has intensified its monitoring of new AI developments to prevent advanced systems from being exploited by military intelligence in nations like China and Russia.
On June 12, Anthropic temporarily disabled its Mythos 5 and Fable 5 models following the export control order. However, the situation evolved on Friday when the US government permitted Anthropic to make Mythos 5 available to select "trusted" US entities, partially reversing the prior directive.
The vetting process implemented by the US government regarding which firms can access these models has faced criticism. OpenAI CEO Sam Altman commented last week, expressing that while extensive safety testing is beneficial, he is not in favor of the government determining which customers should receive access.
At the request of the US government, OpenAI also postponed a broader launch of its GPT-5.6 in order to limit access to a small circle of pre-approved partners.
