Anthropic said on Tuesday, June 30 that the Commerce Department had rescinded export restrictions on its Claude Fable 5 and Mythos 5 models, concluding a tense confrontation with the Trump administration.
In a post on X, Anthropic expressed gratitude to users and collaborators for their patience and assistance in redeploying the models.
The company had cut off access to both models in mid-June following a government directive that invoked "national security authorities." According to Anthropic, the order required halting all access "by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees."
Anthropic co-founder Tom Brown took over negotiations with the administration, stepping in for CEO Dario Amodei. The administration has targeted Amodei due to his vocal stances on AI safety and his public endorsement of Kamala Harris during the 2024 presidential race.
Get your free investing masterclass bonus when you join Market Briefs, our free daily newsletter
On June 26, Howard Lutnick, the U.S. Commerce Secretary, authorized Anthropic to offer Mythos 5 to a limited set of companies and government bodies. In correspondence obtained by CNBC, Lutnick wrote that he had found that proper protections existed to allow some "trusted partners" to use the model.
On Tuesday, June 30, Anthropic confirmed that the full ban on both models had been removed. Lutnick posted on X: "Over the past two weeks, we have worked closely with Anthropic to analyze and approve Fable 5 to ensure alignment across the US Government and strengthen America's leadership in AI."
The government's actions against Anthropic came at a time when Chinese open-source models have been rapidly improving, matching the performance of leading U.S. models while costing significantly less. By restricting Anthropic's model deployment, the Trump administration inadvertently gave Chinese developers a window to close the gap, a concern voiced by many tech executives and investors. The episode has left AI industry leaders puzzled, as they lack clarity on the government's regulatory direction and who wields influence in the White House, especially after David Sacks stepped down as crypto and AI czar earlier this year.
Additionally, the model will be available for up to half of weekly usage quotas until July 7 for subscribers to Pro, Max, Team, and certain enterprise plans, according to Anthropic.
The company will keep collaborating with the U.S. government to broaden Mythos 5 availability for additional domestic and international entities via its Glasswing program, a cybersecurity effort that grants chosen organizations advanced AI models for defensive testing. Additionally, Anthropic will soon make Fable 5 available again via Amazon Web Services, Google Cloud, and the Microsoft Foundry platform.
The Trump administration has indicated a more hands-on approach to AI oversight, with the president signing an executive order on AI in early June. The order, lacking concrete specifics, called on AI developers to voluntarily provide their models for government capability assessments before full release. Federal agencies were given 60 days to create the necessary frameworks and procedures.
This has posed difficulties for firms like OpenAI, which have attempted to launch models during this period. On Friday, OpenAI unveiled three new models and stated it is following the government's directive to initially restrict releases to a "small group of trusted partners." OpenAI noted that it demonstrated the models' capabilities, including GPT-5.6, and briefed the government on its plans prior to the Friday release.
"We don't believe this kind of government access process should become the long-term default," OpenAI said in Friday's post. "It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them."
Subscribe to Market Briefs, our free daily newsletter, and claim your bonus investing masterclass