OpenAI has announced a limited rollout of its new GPT-5.6 AI model family after a request from the U.S. government to restrict access to advanced AI systems.

The company said GPT-5.6 will initially be available only to a small group of trusted partners whose participation has been shared with government officials. The release includes three models: GPT-5.6 Sol, OpenAI’s most advanced model, alongside Terra and Luna, which are designed for different performance and cost requirements.

According to OpenAI, the restriction is a temporary measure while discussions continue around future AI safety and cybersecurity frameworks. However, the company also stated that government review processes should not become the standard approach for future AI model launches.

OpenAI claims GPT-5.6 Sol delivers major improvements in coding, cybersecurity, reasoning, and complex problem-solving. The model includes enhanced security protections designed to reduce misuse and make it more resistant to jailbreak attempts and adversarial attacks.

The announcement comes amid increasing government scrutiny of powerful AI systems and growing debates about how advanced AI models should be regulated before public release.

OpenAI says broader access to GPT-5.6 through ChatGPT, Codex, and its API is expected in the coming weeks.

Read More on VitalStack

Enjoyed this article?

Subscribe for weekly deep-dives on AI and health — straight to your inbox.