OpenAI to launch GPT-5.6 publicly on Thursday

OpenAI will publicly launch its new GPT-5.6 model series on Thursday, after the US government approved a broader release. The rollout includes three models: Sol, Terra and Luna.

OpenAI shared preview access with US-only partners in late June at Washington's request, and the government has now cleared a wider release, Axios reported.

When will GPT-5.6 be released?

OpenAI will release GPT-5.6 to the public on Thursday. The lineup includes Sol, the company's new flagship model, Terra, a mid-range option built for everyday work, and Luna, a fast and low-cost model. OpenAI said in a Tuesday post on X that it is also expanding preview access globally.

Why was GPT-5.6's launch delayed?

OpenAI limited GPT-5.6 preview access to a small group of US-only partners in late June, at Washington's request. The restriction followed concern over the model's supposedly unprecedented ability to identify software vulnerabilities, the weaknesses in code that hackers exploit. The Trump administration has since approved a broader launch, following testing and meetings between officials and the company.

Axios reported Tuesday, citing a source familiar with the matter, that the administration gave OpenAI the green light for the wider GPT-5.6 rollout. AFP said it had contacted OpenAI, the White House and the US Department of Commerce for comment on the report.

Is Anthropic facing a similar situation?

Anthropic dealt with a comparable restriction on its own advanced models, Fable 5 and Mythos 5. Last week, the company said it would begin restoring global access to both models after the US government lifted a limit on where they could be released. The parallel moves at OpenAI and Anthropic point to shifting export rules around the most capable AI systems.