OpenAI to Release GPT-5.6 Models After U.S. Government Review

OpenAI will publicly release its GPT-5.6 Sol, Terra and Luna models on Thursday after the U.S. government completed additional testing and meetings with the company. The company announced the launch late Tuesday on X. Officials had previously asked for the delay over national security concerns.

Trump signed an executive order creating a voluntary framework. Under the order, developers can provide covered frontier models to the U.S. government for up to 30 days before releasing them to trusted partners. OpenAI had previously limited access to a small group of vetted partners whose details were shared with authorities.

Anthropic faced parallel restrictions.

The company disabled Mythos 5 and Fable 5 for all users after a U.S. export control order on June 12. Controls on Fable were lifted last week after safeguards were added. Mythos remains available only to some trusted U.S. organizations.

Musk announced Wednesday that SpaceXAI is making Grok 4.5 available to the public. OpenAI had previewed the GPT-5.6 models in late June and highlighted improved agentic capabilities in coding, biology, and cybersecurity.