GPT-5.6 rollout limit: Trump administration asks OpenAI to restrict initial release
The Trump administration asked OpenAI to limit the initial release of GPT-5.6 to a small group of government-approved partners while the model undergoes evaluation. The request, made by the White House’s Office of the National Cyber Director and the Office of Science and Technology Policy, is the second U.S. government intervention this month after an order directed Anthropic to suspend public access to Claude Fable 5 and Mythos 5 over national security concerns. The matter touches companies including OpenAI, Anthropic and Google.
The White House’s Office of the National Cyber Director and the Office of Science and Technology Policy made a formal request to OpenAI to limit the initial release of GPT-5.6 to a small group of government-approved partners while the model is evaluated. The request identified those two White House offices as the proponents of the restriction. It asked that access be limited during the evaluation period.
The motivation for the GPT-5.6 rollout limit was to provide time for the administration to develop a framework for evaluating advanced AI models prior to wider deployment. The request aims to limit rollout while that framework is developed. Reports said the request was driven by GPT-5.6’s “Mythos-like” capabilities. Those capabilities were cited in prompting the evaluation-focused limitation.
This summary presents the offices involved and the motivations cited in the request. It does not detail mechanisms or specific policy steps.
The executive order earlier this month directed federal agencies to establish a voluntary testing framework for advanced AI systems before release. While that framework is being developed, the White House’s Office of the National Cyber Director and the Office of Science and Technology Policy made a request to OpenAI to limit the initial release of GPT-5.6 to a small group of government-approved partners during an evaluation period. The request aims to limit rollout while the administration develops a framework for evaluating advanced AI models prior to wider deployment. The request was reportedly driven by GPT-5.6’s “Mythos-like” capabilities.
Anthropic, OpenAI, and Google have published proposals outlining governance for frontier AI, including structured evaluations, greater safety testing transparency, independent review, and increased government oversight. These proposals set out measures intended to govern advanced AI systems before broad deployment. The published proposals originate from the three companies named in the provided facts.
The administration’s intervention may test whether governance frameworks can be applied evenly across the industry, with concerns about potential regulatory capture. Sam Altman’s 2023 Senate testimony advocated for a regulatory agency for advanced AI systems. Dario Amodei argued that rigorous government-backed evaluations are needed due to national security risks.
The Trump administration’s request to limit the GPT-5.6 rollout and related interventions involving frontier AI models represent a heightened level of government involvement in the oversight of advanced artificial intelligence deployments. These efforts have followed a cautious, policy-focused trajectory that emphasizes evaluation, governance and measured oversight while broader frameworks for assessing and managing such systems are developed, refined and coordinated.
