OpenAI GPT-5.6 Launch: US Gov. Limits Access Over Cyber Fears

OpenAI has released GPT-5.6 with three models: Sol, Terra, and Luna. Access is immediately restricted to about 20 companies. This limited rollout follows a request from the US government, citing national security concerns. TechCrunch confirmed this restricted release. This marks the first time a US administration has asked an American AI company to limit model availability before a broad launch. OpenAI expects to expand access within weeks, with a full public rollout to follow. Sol is designed for frontier reasoning. Terra offers a balance of performance and cost, competitive with GPT-5.5 at about half the price. Luna is the speed-focused option, built for affordability. The government's request is linked to the model's advanced cyber capabilities. OpenAI states GPT-5.6 Sol does not reach a "critical" level in its preparedness framework. It says the model is better at finding vulnerabilities than carrying out attacks. This restricted rollout is a test case for assessing frontier AI models' cyber capabilities. For developers, this precedent suggests future AI releases might face similar friction due to cybersecurity concerns.