OpenAI has initiated a limited US preview of its GPT-5.6 series – Sol, Terra, and Luna – following a request from the United States government. This move occurs after the US restricted foreign access to rival Anthropic’s AI models, citing national security concerns related to their ability to identify software vulnerabilities. OpenAI briefed officials on the new models’ capabilities before launch and is limiting initial access to trusted, US-based partners, including those with overseas employees. Terra, the mid-range model, will be priced at half the cost of its predecessor, GPT-5.5, as OpenAI aims to maintain market share against competitors like Anthropic and Google. The government’s intervention is notable given its recent efforts to reduce AI regulation, but was prompted by the advanced capabilities of these new models. A recent executive order now requires voluntary federal review of national security risks associated with advanced AI before release.