OpenAI is limiting access to its newest artificial intelligence model at the request of the US government, the company said Friday in a blog post.
The restricted release covers OpenAI's freshly announced GPT-5.6 lineup, a trio of celestial-themed models named Sol, Terra and Luna. According to The Hindu Business Line, the company is holding back broad access to GPT-5.6 Sol while a US government cybersecurity review plays out.
The move follows a request tied to the Trump administration. According to CBS8, OpenAI restricted the new GPT release at the administration's request. The Washington Post reports that OpenAI is limiting its latest ChatGPT product to Trump-approved customers during the cybersecurity review, meaning access is being granted selectively rather than to the general public.
The situation marks an unusual moment: a leading AI developer announces a flagship model, then pauses its wider rollout because a government has asked it to. As The Hindu Business Line notes, the episode is fueling debate over AI oversight, cybersecurity risks and the pace of innovation. One outlet framed it bluntly, saying OpenAI wants to drop its next model but the US government has stepped in to put a hand on the brake.
The available reporting does not spell out exactly what security concerns prompted the review, how long it will last, or which customers qualify for early access. What is clear is that the restriction came at the government's request and is tied to a cybersecurity assessment, per the company's own blog post and multiple outlets.
Why it matters: when a government can slow the public release of a powerful new AI system, it signals that frontier models are increasingly being treated as matters of national security — a shift that could reshape who gets access to the most advanced AI, and when.