Source: CNN — The White House has requested OpenAI limit the release of its upcoming GPT-5.6 model to a small number of government-approved partners, a source familiar with the discussions told CNN on June 25. The request centres on the model's advanced capabilities, which the administration considers sensitive enough to warrant restricted distribution.
The development mirrors the Trump administration's June 12 order that forced Anthropic to withdraw its Mythos 5 and Fable 5 models — an export control directive that subsequently led to the US allowing a partial redeployment of Mythos 5 to roughly 100 trusted cybersecurity firms on June 27. OpenAI's situation appears to be moving in the same direction: pre-emptive restrictions rather than post-release pullbacks.
Neither OpenAI nor the White House have publicly commented on the specifics of the request, but the pattern is unmistakable. Frontier AI models are now being treated as dual-use technology subject to national security review, a designation that would fundamentally reshape how AI companies release products.
Analysis
The White House isn't asking OpenAI to limit GPT-5.6 because it's worried about deepfakes or misinformation. It's worried about foreign intelligence services fine-tuning frontier models on their own classified datasets — something that becomes dramatically easier when the base model is openly available. The timing is the real story: this is a pre-emptive restriction, not a reactive one. The administration learned from the Anthropic scramble, where it had to retroactively force a pullback and then negotiate a carve-out for critical infrastructure firms. With OpenAI, it's trying to shape the release before it happens. The unanswered question is whether OpenAI will comply quietly — and what that compliance buys them in the IPO narrative they're now constructing for 2027. Cooperation on national security might be a prerequisite for a smooth public offering.
