OpenAI will initially only release ChatGPT 5.6 to government-approved customers

At a glance:

• OpenAI will limit initial ChatGPT 5.6 access to U.S. government‑approved customers.
• The staggered rollout follows a staff memo from CEO Sam Altman indicating a broader release a couple of weeks later.
• The move comes after an executive order from President Trump urging voluntary federal review of powerful AI models.

What happened

OpenAI informed employees via a memo from CEO Sam Altman that the first users of ChatGPT 5.6 will be parties approved by the federal government, with access granted "approving access customer by customer during this preview period," according to the memo. The memo also quoted Altman saying, "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."

Several U.S. agencies are involved in shaping this approach, including the Office of the National Cyber Director, the Office of Science and Technology Policy, and the Department of Commerce under Secretary Howard Lutnick. Neither the White House nor the Office of the National Cyber Director responded to requests for comment from The Information.

The report notes that shortly after Trump’s executive order, rival Anthropic disabled access to two of its recent models following a federal directive that sought to block foreign nationals from using its tools, highlighting the unclear and evolving nature of the voluntary review process.

Why it matters

This selective release signals a shift toward greater government oversight of cutting‑edge AI models, potentially setting a precedent for how companies handle pre‑release scrutiny. Developers and enterprise customers may face delays in accessing the latest capabilities, which could affect product timelines and competitive dynamics.

By framing the rollout as a voluntary preview, OpenAI attempts to balance cooperation with regulators while preserving its commercial roadmap, but the lack of a clear, standardized framework leaves uncertainty about how long the government‑approved phase will last and what criteria will be used for approval.

Looking ahead, the industry will watch whether the promised "couple of weeks later" broader rollout materializes, how the federal review framework evolves, and whether other AI firms adopt similar staggered releases in response to increasing regulatory pressure.