The Trump administration is effectively picking winners for access to frontier AI models, with Anthropic and OpenAI both bending to government demands this week.
Commerce Secretary Howard Lutnick notified Anthropic on Friday that Claude Mythos 5, the company's most powerful cybersecurity model, can be redeployed to more than 100 US organizations including major companies and federal agencies. The letter, addressed to Anthropic co-founder Tom Brown, stated Lutnick had "determined that appropriate safeguards are in place to permit certain trusted partners to access" the model. The partial approval ends a two-week standoff.
On June 12, the government ordered Anthropic to block all foreign national access to Mythos 5 and its consumer-grade sibling Fable 5, forcing the company to disable both models entirely. The directive even covered foreign national Anthropic employees working inside the United States. The government stopped short of allowing Fable 5's return.
Lutnick's letter is "silent on Fable 5," Semafor reported, though people close to the talks said discussions are moving toward releasing it as well. The same day, OpenAI announced it was delaying broad access to its GPT-5.6 models at the US government's request, limiting the initial rollout to a "small group of trusted partners" approved in coordination with the administration. OpenAI said it previewed the models' capabilities to the government ahead of Friday's launch.
"We don't believe this kind of government access process should become the long-term default," OpenAI said in a statement. "It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them." The back-to-back actions mark a sharp turn for the Trump administration, which signed an executive order on June 2 taking a light-touch approach to AI regulation.
That order let model makers voluntarily submit advanced systems for federal review up to 30 days before public release. The clampdown on Anthropic and the voluntary limits on OpenAI suggest a more interventionist posture is emerging without legislation.
Anthropic's relationship with the administration has been particularly fraught. Earlier this year, the Defense Department labeled the company a "supply chain risk" after negotiations over military use collapsed.
Anthropic sued over the designation and has notched an early win in the ongoing case. The White House grew concerned about Mythos after learning Anthropic granted access to a South Korean telecom firm it believed had ties to China, WIRED reported.
Amazon and the NSA separately flagged a potential jailbreak of Fable 5. The confluence pushed officials to act.
Anthropic confirmed it is restoring access for approved organizations and continuing talks with the White House about Fable 5 over the weekend. The company said it is "pleased to see this progress."
Investor David Sacks, speaking about Anthropic CEO Dario Amodei's April warning about Mythos, said the CEO "spiked the cortisol level, got everyone really worried. And there was some truth to it in terms of the sense that this model had advanced cyber capabilities."
