OpenAI limits GPT-5.6 rollout after government request, says restrictions shouldn’t be the norm
OpenAI is restricting the release of its new GPT-5.6 models to a small group of trusted partners at the U.S. government's request, but the company argues this should not become a long-term practice.
OpenAI announced on Friday that it is limiting the release of its latest AI models to a “small group of trusted partners” in response to a request from the U.S. government. The new GPT-5.6 lineup includes three models: Sol (flagship), Terra (balanced for everyday use), and Luna (fast and low-cost). All three are subject to restrictions.
The government’s request comes amid increased pressure on AI companies to restrict their most advanced systems. Earlier, after Anthropic released its powerful Fable 5 model, the administration ordered it to block access for foreign nationals, leading Anthropic to take the model down entirely.
Dean Ball, a former White House AI advisor who will soon join OpenAI, said President Trump’s recent executive order—which asks AI companies to voluntarily submit advanced models for government review up to 30 days before release—has created a de facto involuntary licensing regime. Ball argued that without clear safety standards, this could lead to endless delays, benefiting China in the AI race and jeopardizing billions in AI infrastructure investments.
OpenAI made it clear it was not happy with the arrangement. “We don’t believe this kind of government access process should become the long-term default,” the company wrote in a blog post. “It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them.” OpenAI called the limited preview a “short-term step” that will pave the way for broader availability in the coming weeks as it works with the administration on a new cybersecurity executive order framework and a repeatable process for future releases.
GPT-5.6 Sol is OpenAI’s strongest model, featuring improved agentic capabilities in coding, biology, and cybersecurity. It introduces a “max” reasoning effort mode and an “ultra” mode that uses coordinated subagents for complex tasks. OpenAI says Sol slightly outperforms Anthropic’s Claude Mythos 5 in coding workflows while using a third of the output tokens. Sol also includes enhanced safety guardrails built directly into the core model, designed to be hard to jailbreak and favoring defensive cybersecurity. This approach aims to avoid the over-cautious routing issues that plagued Anthropic’s Fable 5.
Initially available only to select partners, OpenAI plans to roll out GPT-5.6 models more broadly via ChatGPT, Codex, and the API soon. Pricing: Sol costs $5 per million input tokens and $30 per million output tokens; Terra costs half that; Luna costs $1 and $6, respectively. Improved prompt caching is also included to reduce costs for repeated queries.
