OpenAI gets Commerce clearance for GPT-5.6 release, Axios says

The U.S. Department of Commerce has cleared OpenAI to move ahead with a broad release of its GPT-5.6 model, Axios reported, citing a person familiar with the matter. The decision could allow OpenAI to expand access as soon as this week, reducing a period of regulatory uncertainty around frontier AI releases in the United States.

Axios reported that the rollout is expected after further testing and meetings between OpenAI and government officials. CNBC said OpenAI, the White House and the Commerce Department did not immediately respond to its requests for comment.

The reported approval follows OpenAI’s unveiling late last month of three models: GPT-5.6 Sol, Terra and Luna. At that time, the company limited availability to a small group of trusted partners as part of a phased release intended to comply with federal oversight, according to CNBC. OpenAI did not identify those partners.

In a company blog post cited by CNBC, OpenAI said it supports broad access and was working to make the models generally available in the coming weeks. The reported Commerce clearance would mark a step toward that wider distribution, at least for GPT-5.6.

How the release process works

A phased launch gives a model developer time to test systems with a limited set of users before wider availability. In this case, OpenAI initially kept access to selected partners while it addressed federal oversight requirements, according to CNBC. Axios reported that additional testing and discussions with officials remained part of the process before the broader rollout.

The approach reflects a more active role for Washington in reviewing advanced AI systems before they reach a wider market. CNBC reported that the Trump administration has taken a more hands-on stance toward AI regulation, with an effort to assess model capabilities before full-scale release.

Export controls can affect AI model access by limiting where and how advanced systems are made available. CNBC reported that Anthropic, another U.S. AI company, had suspended access to its Claude Fable 5 and Mythos 5 models last month while complying with government export controls. The restrictions on those models were lifted last week, CNBC reported, ending a period in which availability was constrained for users globally.

Competitive pressure from China

The tighter U.S. review process has also had competitive implications. CNBC reported that limits on domestic frontier AI releases have created an opening for Chinese competitors offering more accessible and lower-cost models.

One example is Zhipu, which trades as Knowledge Atlas Technology JSC. CNBC reported that the company launched its GLM 5.2 model last month. The model is free to download, fine-tune and run on an enterprise’s own servers, according to CNBC.

For OpenAI, the reported Commerce decision removes one barrier to broader GPT-5.6 availability. The timing and scope of the rollout remain tied to the company’s testing process and its meetings with U.S. officials, according to Axios.

This story draws on original reporting from CNBC.