TL;DR
The Trump administration is dictating access to cutting-edge AI models, moving that decision from Anthropic and OpenAI to the government through the Gold Eagle program.
The Trump administration is now dictating which companies and entities have access to Anthropic and OpenAI’s cutting-edge artificial intelligence models, CNBC reported Friday, citing two people familiar with the matter. Until now, that decision was made by the laboratories themselves. Anthropic controlled access to its Mythos cybersecurity model through an initiative called Project Glasswing. OpenAI ran a similar program called Daybreak for its cybernetic model. In the future, these partner lists will require explicit government approval.
A White House official he told CNBC that the government does notprovide approvals for AI releases” and that the company’s participation is “volunteer.“But the administration blocked Anthropic’s Claude Mythos 5 and Fable 5 last month on national security grounds, restoring access only after weeks of negotiations. OpenAI said in June it would limit new models to “trusted partners“To comply with government requests. The gap between the official position and operational reality is the story. The White House launched Gold Eagle this weeka clearinghouse on cyber AI vulnerabilities and, according to CNBC’s source, the program will put the White House in charge of giving the green light to which companies can access new AI models.
The moment is politically uncomfortable. Moonshot AI’s Kimi K3 launched on the same day and matched or surpassed Fable and GPT-5.6 in at least one independent benchmark. David Sacks, former White House AI czar, called it “about” and wrote: “This is how you lose the AI race. The rest of the world won’t follow our rules if we stagnate.The administration is trying to protect frontier AI from Chinese exploitation while watching Chinese labs close the capability gap in real time.
The change is structural, not temporary. Trump’s executive order in June called on AI companies to give the government early access to models for testing.framed as a volunteer. Gold Eagle operationalizes that question into something more like a gate mechanism. If Anthropic and OpenAI cannot launch their most capable models without government approval from the list of partners, then the US government has acquired de facto distribution authority over frontier AI, without legislation, without a regulatory agency and through a program that the White House insists is optional.
