An article summarized by Axios:

The Trump administration has approved OpenAI to broadly release its latest flagship AI model, GPT-5.6, after additional government testing and discussions with federal officials. OpenAI announced that GPT-5.6, known as Sol, along with its smaller Terra and Luna models, will become publicly available on Thursday.

The approval follows a staggered rollout that began last month, when the administration limited access to government-approved organizations while the model underwent further evaluation. Testing was conducted by the Department of Commerce's Center for AI Standards and Innovation, with OpenAI engineers working directly with government officials to answer technical questions. OpenAI has said it would have preferred a full public release from the start.

The decision highlights the growing role the U.S. government is playing in overseeing the release of advanced AI systems. Both OpenAI and Anthropic have recently faced restrictions on deploying their most powerful models, with Anthropic's Mythos and Fable models temporarily limited by the Commerce Department before access to Fable was restored last week. The case underscores how AI companies and regulators are shaping access to cutting-edge technology in real time as broader safety standards are still being developed.

For the article: https://www.axios.com/2026/07/08/openai-gpt-trump-ban-lifted