The Trump administration is now dictating access to frontier AI models

The Trump administration has taken new steps to take more control over future AI model releases by determining which companies and organizations will be allowed access to the latest edge models, two sources familiar with the matter told CNBC.

Until now, this decision was in the hands of American artificial intelligence giants.

both anthropic And OpenAI They decided which companies and agencies had access to the most powerful models, and often included large corporate clients.

Anthropic has introduced its most capable Mythos cybersecurity model to a handful of partners with Project Glasswing. OpenAI was asked by management to approve the latest GPT-5.6 release and has a similar consortium called Daybreak for its cybersecurity model.

A White House official told CNBC that private companies do not approve AI broadcasts.

The official told CNBC that any interviews, tests or meetings with government experts were “voluntary” and that “decisions regarding the timing and scope of releases rest entirely with the companies,” citing Trump’s recent executive order.

“The administration continues to collaborate with all of America’s frontier laboratories to strengthen the security of this technology without stifling innovation,” they wrote.

But last month, the Trump administration blocked Claude Mythos 5 and Fable 5 over “national security concerns” and restored access after weeks of intense negotiations with Anthropic. Last month, OpenAI said it would limit new AI models to “trusted partners” to comply with government demands.

The White House is walking a fine line on regulation at a time when advanced AI tools pose major cybersecurity risks and cheaper, vulnerability-heavy models from China are rapidly closing the gap with American border labs.

Chinese startup Moonshot AI introduced the Kimi K3 model on Friday, which largely matches the performance of Fable and GPT-5.6 and even outperforms the US border models in at least one independent benchmark.

David SackThe founder of Craft Ventures and former White House AI czar called Kimi’s breakthrough “concerning.”

“This is how you lose the AI ​​race,” he wrote. “If we get ourselves bogged down, the rest of the world won’t play by our rules.”

The administration has taken several steps in recent months to reshape AI oversight, starting with President Donald Trump’s June executive order asking companies to voluntarily give the government early access to models for testing.

This week, the administration launched its own program called “Golden Eagle” aimed at collaborating with the private sector to find and fix cybersecurity vulnerabilities.

The so-called clearinghouse will hold the White House responsible for giving the green light for companies to access new artificial intelligence models, according to a source familiar with the matter who spoke on condition of anonymity to discuss nonpublic information.

The administration’s moves have thrown the future of company-led initiatives like Project Glasswing and OpenAI’s Daybreak into doubt.

Going forward, these deployments will require explicit government approval on which partners can be involved, according to a source.

— CNBC’s Megan Cassella and Ashley Capoot contributed reporting