OpenAI’s next major model, ChatGPT 5.6, won’t be available to everyone right away. The company plans to release it first to a select group of users who have been individually approved by the federal government, with a broader rollout expected only a couple of weeks after that initial preview period.
According to Engadget, citing a report from The Information, a staff memo from CEO Sam Altman confirmed the approach. Federal leaders will be “approving access customer by customer during this preview period,” Altman reportedly wrote, before a wider release of the 5.6 model follows “a couple of weeks later.”
Altman also reportedly acknowledged this isn’t the direction he wants to go long term. “We’ve made clear to the US government that this is not our preferred long term model, and will work with them and others in industry to achieve a more sustainable approach for future releases,” he told employees in the memo.
Several government bodies appear to be involved in shaping this rollout. The Information named the Office of the National Cyber Director and the Office of Science and Technology Policy, with Department of Commerce Secretary Howard Lutnick also reportedly playing a role. Neither the White House nor the Office of the National Cyber Director responded to the publication’s requests for comment.
The context here matters. Earlier this month, President Donald Trump signed an executive order asking AI companies to voluntarily submit their more powerful models for federal review before public release. The government is still working on a standardized framework for how that review process will actually work.
But “voluntary” is starting to look like a flexible term. Shortly after that executive order, OpenAI rival Anthropic was forced to disable access to two of its recent models following a federal directive. That order gave no specific reasons for the security concerns, only that the government wanted to cut off access for foreign nationals. The details were thin, and the reasoning was unclear.
Taken together, these two incidents point to something significant happening beneath the surface of U.S. AI policy. The gap between what was promised, a light-touch voluntary review process, and what is actually happening, government officials approving users one by one and blocking access to models, is wide. Companies like OpenAI and Anthropic are navigating pressure that appears to be more directive than the original executive order implied.
For everyday users and businesses building on top of these models, the immediate concern is access. If you were waiting for ChatGPT 5.6, you may have to wait longer than expected unless your organization is already on a government-approved list. For the AI industry more broadly, the bigger question is what this sets as a precedent. If federal approval becomes a standard step in model releases, that changes the speed at which new AI tools reach the market and who controls that timeline.
