OpenAI will stagger the release of GPT-5.6, restricting initial access to customers approved by the federal government. CEO Sam Altman told staff in an internal memo that federal leaders will be “approving access customer by customer during this preview period,” with a broader public release expected “a couple of weeks later,” according to The Information as reported by Engadget.
The arrangement marks the first time the US government has preemptively gated the release of an American AI model before public availability.
The Agencies Involved
Multiple federal agencies are directing the process. Engadget identified the Office of the National Cyber Director and the Office of Science and Technology Policy as key participants, with Commerce Secretary Howard Lutnick also involved. Neither the White House nor the Office of the National Cyber Director responded to requests for comment.
Altman pushed back on the long-term implications. “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 reportedly told employees, according to Engadget.
A Pattern Forming
The GPT-5.6 restriction follows a broader policy shift. President Trump signed an executive order earlier this month asking AI companies to participate in a voluntary federal review of their more powerful models before public release, with a framework for standardized assessment reportedly in development.
That “voluntary” framing is already showing cracks. Anthropic recently disabled access to two of its models following a separate federal directive from the Commerce Department that blocked foreign nationals from using Anthropic’s tools, Engadget reported. Between that directive and the GPT-5.6 gating, the boundary between voluntary review and mandatory compliance is blurring.
The Deployment Gate
For teams building on frontier models, the practical question is timing and access. If GPT-5.6 includes the multi-step reasoning and tool-use capabilities expected at this model generation, agent developers relying on OpenAI’s API face a period where their toolchain depends on government approval structures rather than release schedules.
Enterprise buyers in regulated industries (finance, healthcare, defense) may benefit from the approval process as a compliance signal. Smaller developers and startups face the opposite problem: waiting weeks for general availability while competitors with government relationships get early access.
Altman’s memo suggests he views the arrangement as temporary. Whether the “more sustainable approach” he referenced means reverting to self-governed releases or formalizing a permanent review process will likely shape how every frontier AI company plans model launches going forward.