OpenAI Launches “Next-Gen” GPT‑5.6 Models in Limited Preview

OpenAI just announced the release of three next-generation frontier models, GPT-5.6 Sol, Terra, and Luna. The three models are available as a limited preview in Codex and the API for a small group of partners, and the AI startup said that it’s proceeding that way at the request of the US government.

According to OpenAI, GPT-5.6 Sol is its “most capable model yet” for cybersecurity, being a close contender to Anthropic’s Mythos Preview on the ExploitBench API, all while “using only ~1/3 of the output tokens.” GPT-5.6 Sol also launches with OpenAI’s “most robust safety stack yet,” with “strengthened protections for higher-risk activity, sensitive cyber requests, and repeated misuse.”

With GPT-5.6, OpenAI also promises “state of the art” performance for complex command-line workflows, with the new model setting a new record on the TerminalBench 2.1 benchmark for scientific computing in terminal environments. GPT-5.6 Terra, however, is said to be offering a GPT-5.5 level of performance at 2x lower cost. The third GPT-5.6 model, Luna, is the most efficient and affordable one.

We believe in broad access and plan to make GPT-5.6 Sol, Terra, and Luna generally available in the coming weeks.

For now, at the request of the U.S. government, we’re starting with a limited preview among a small group of trusted partners in Codex and the API.

— OpenAI (@OpenAI) June 26, 2026

After the US government recently forced OpenAI rival Anthropic to disable its most advanced AI models to limit foreign access, OpenAI said that it previewed the capabilities of its latest flagship models with the government. However, the AI startup expressed its disagreement with how the US government is addressing risks associated with powerful AI models.

“We don’t believe this kind of government access process should become the long-term default. It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them. We are taking this short-term step because we believe it is the strongest path to broader availability in the coming weeks, while we work with the Administration to develop the cyber Executive Order framework and a repeatable process for future model releases,” OpenAI said today.