The Trump administration has asked OpenAI to restrict the initial release of its upcoming model, GPT-5.6, to a select group of government-approved partners due to security concerns. This unprecedented request marks the first time the U.S. government has preemptively sought to limit an American AI company's model launch. The White House, specifically the Office of the National Cyber Director and the Office of Science and Technology Policy, made the request while developing a framework for evaluating new model security. OpenAI CEO Sam Altman, while acknowledging the limited rollout plan in an internal memo, stated it's not their preferred long-term approach. OpenAI has been collaborating with the administration on the model release, and the White House has even previewed GPT-5.6's capabilities. Altman discussed the model with Commerce Secretary Howard Lutnick, who emphasized government-wide testing and approval. The government's intervention stems from GPT-5.6's "Mythos-like" capability, indicating its high power and the administration's desire for robust safeguards. This move aligns with a recent executive order signed by President Trump, which mandates voluntary testing protocols for AI companies before new model releases. AI labs face a dilemma: competing to release new models while security officials worry about misuse by malicious actors. Altman hopes for a wider release of GPT-5.6 within a "couple of weeks" after the initial restricted rollout.