OpenAI will release GPT-5.6 in limited preview form to enterprise customers only, following a request from the Trump administration citing security concerns. CEO Sam Altman announced the staggered rollout to employees Wednesday.
The Trump administration has asked OpenAI to delay the full release of GPT-5.6, its next major AI model, according to reporting from The Information. The request stems from apprehension over potential security risks.
Under the new plan, GPT-5.6 will initially be available only to a limited group of enterprise customers rather than a broad public launch. During this preview period, the Trump administration will have access to evaluate the model before wider deployment.
Sam Altman disclosed the decision during a company Q&A with employees Wednesday, framing the delay as compliance with a federal government request.
The move follows similar regulatory pressure on competitors. Anthropic suspended its most capable AI offerings from the market nearly two weeks ago under pressure from regulators.
The staggered release strategy represents a shift in how advanced AI models reach the market. Rather than immediate public availability, the approach gives government entities time to assess security implications before broader deployment.
OpenAI has not specified a timeline for when GPT-5.6 will become widely available beyond the initial enterprise preview phase. The company did not immediately respond to requests for additional details on the security concerns cited by the administration or the scope of the initial preview group.
The request underscores growing government attention to large language model releases and reflects broader concerns about AI safety and security at the federal level. It also signals the Trump administration's willingness to directly influence the timeline and rollout strategy of major AI systems from leading companies.
The Pentagon has revised its military doctrine to allow AI systems to initiate targeting actions with human oversight, marking a significant shift in automated warfare protocols.
Notion is discontinuing Notion Mail on September 22, pivoting instead toward AI agents to manage inboxes. The company cited low engagement, with over 50% of users never opening the inbox.
The Pentagon has revised its military doctrine to allow artificial intelligence greater authority in selecting combat targets. The policy shift enables AI systems to make critical wartime decisions previously reserved for human commanders.
Global AI revenue excluding China reached $25 billion in the first quarter, marginally exceeding the estimated $21 billion in data center and chip depreciation costs. The narrow margin underscores persistent profitability challenges in the sector.