What happened
The White House reportedly requested OpenAI restrict the initial release of its GPT-5.6 model. OpenAI plans a restricted preview program, making GPT-5.6 available only to a small group of approved customers, per The Information. OpenAI CEO Sam Altman reportedly confirmed government approval for customer access during this phase, with wider availability expected "a couple of weeks later." This follows similar reported restrictions on Anthropic's Mythos and Fable AI models.
Why it matters
Government oversight now directly shapes frontier AI model deployment, shifting from broad public launches to controlled, government-supervised previews. This mechanism restricts initial access to advanced capabilities, impacting procurement teams evaluating new AI systems and security architects assessing integration risks. The reported restrictions on GPT-5.6, following similar limitations on Anthropic's cyber-focused Mythos and Fable models, establish a precedent for national security considerations overriding rapid commercial availability for highly capable AI. Founders developing AI-powered solutions must now factor in potential government-mandated release delays and access controls.
