The Trump administration has asked OpenAI to restrict the general release of its forthcoming GPT-5.6 AI model, instead allowing access only to a limited set of government-approved partners initially. This request stems from concerns regarding the model's advanced capabilities, which are considered to be "Mythos-like," referencing Anthropic's highly capable and now restricted models (Mythos 5 and Fable 5). The government's intervention underlines a growing apprehension about the potential misuse of powerful AI by malicious actors, including nation-state spies and cybercriminals, who could exploit software vulnerabilities at speeds unmatched by human analysts.
OpenAI CEO Sam Altman confirmed this limited rollout in an internal memo to employees, stating that the government would be "approving access customer by customer" during this preview period. Altman also communicated that this is not OpenAI's preferred long-term model for releases and that they aim for a broader public release "a couple of weeks later," provided the limited rollout proves successful. This proactive engagement from OpenAI with the administration began before the Commerce Department issued a rare directive restricting Anthropic's models.
The White House's Office of the National Cyber Director and Office of Science and Technology Policy were instrumental in requesting this limited release, as the administration is actively developing a framework for testing and evaluating the security of new AI models. This action follows President Trump's recent executive order on AI, which encouraged voluntary submission of new models for government review before public release. The situation highlights the current lack of a clear, consistent federal regulatory framework for AI, leading to confusion among AI companies regarding oversight responsibilities, with requests originating from different government entities.
