The U.S. government has lifted export controls on Anthropic's flagship large model, ending a weeks-long standoff between the artificial intelligence company valued at $1 trillion and the Trump administration. According to people familiar with the matter, the U.S. Department of Commerce notified Anthropic on Tuesday night to lift the ban on overseas users accessing the company's Mythos and Fable series models. The resolution allows the artificial intelligence company to reopen its latest model, Fable5, to the public.
Last week, the U.S. government gave Anthropic approval to relaunch Mythos 5 to about 100 pre-vetted partner companies. This model has fewer security protection mechanisms and is mainly aimed at enterprise customers.
A letter obtained by the media from Commerce Secretary Howard Lutnick to Anthropic co-founder Tom Brown shows that the premise for lifting the ban is that Anthropic "commits to proactively identify and resolve security risks related to its models."
A person close to Anthropic said that the company has implemented a new security protection mechanism to specifically address the risks related to the ban. The person added that this protection solution has passed testing and approval by the U.S. government’s Artificial Intelligence Standards and Innovation Center.
OpenAI’s new model GPT-5.6 also encountered regulatory restrictions. The San Francisco-based company received a request from the Trump administration to initially only open access to the model to about two dozen officially licensed partners.
A person familiar with the matter revealed that GPT-5.6 will be available to a wider range of users as soon as next week.
At the same time, the United States and Europe have discussed setting up a "trusted partner" mechanism for cutting-edge artificial intelligence models, according to which core U.S. allies can gain priority access to the latest AI tools.