OpenAI has restricted access to its latest suite of advanced models, GPT-5.6, following a request from the U.S. government. The new lineup includes Sol, Terra and Luna, designed for various uses, but their release is limited to select partners vetted by officials.

This move comes as AI firms face increasing pressure from governments to limit access to their most advanced systems. After Anthropic faced similar restrictions on Fable 5, OpenAI finds itself in a delicate position – complying with government requests while questioning the long-term implications of such measures.

Dean Ball, a former White House adviser and now an OpenAI employee, argues that this approach could hinder innovation and push global tech leadership towards competitors. While OpenAI promises to work on new frameworks for model releases, the current process seems heavy-handed and potentially risky.

GPT-5.6 Sol boasts improved agentic capabilities in coding, biology, and cybersecurity. Though robust security measures are in place, concerns remain about government oversight stifling progress. Prices start at $1 per million tokens for Luna, with Sol being the priciest option.