The Trump administration has given OpenAI permission to launch its advanced GPT-5.6 AI model for the public. OpenAI announced that its new AI models – GPT-5.6 Sol (flagship model), Terra, and Luna – will be launched publicly on Thursday. This approval is important because the US government is now closely checking powerful AI models before they are released to everyone.
The decision came after OpenAI completed extra safety testing and held several meetings with US government officials. The testing was carried out by the Center for AI Standards and Innovation, which is part of the US Department of Commerce. According to Axios, OpenAI also sent technical experts to Washington, D.C., where they stayed to answer any questions from government officials during the review process.
GPT-5.6 phased rollout
Last month, the Trump administration did not allow OpenAI to release GPT-5.6 to everyone immediately, as per multiple reports. Instead, it asked the company to launch the models in stages. During that first stage, only government-approved organisations and a small group of trusted partners were allowed to use the models.
However, Axios reported, OpenAI had said at that time that a phased rollout was not its preferred way of launching new AI models. The company also explained that AI companies and the government are still working without final rules for releasing advanced AI systems. OpenAI said those future rules are expected to come under President Donald Trump’s latest AI executive order.
OpenAI broad AI access
OpenAI expects the broader rollout to begin as early as this week after completing the required testing and government discussions, according to CNBC. OpenAI, the White House, and the US Department of Commerce did not immediately respond to requests for comment on the report.
Before this wider launch, OpenAI had announced three new models GPT-5.6 Sol, Terra, and Luna. OpenAI said it “believes in broad access” and that it is starting with “a limited preview for a small group of trusted partners” at the U.S. government’s request before making the models generally available, according to its official post.
Also read: Microsoft replaces OpenAI, Anthropic with its own AI models in Excel, Outlook
OpenAI did not reveal the names of those trusted partners. In its earlier blog post, OpenAI said it believes in giving broad access to its AI models and planned to make them generally available in the coming weeks. The reported approval shows that the Trump administration is taking a more active role in regulating AI by checking powerful models before they are released widely.
Anthropic AI restrictions
OpenAI is not the only AI company facing government restrictions. Rival AI company Anthropic has also recently faced similar limits. In June, the US Department of Commerce banned foreign users from accessing Anthropic’s Mythos and Fable AI models, forcing the company to withdraw them from the market, according to Axios. Anthropic’s Claude Fable 5 and Mythos 5 models were also suspended last month as the company complied with US export control rules.
The restrictions on Anthropic’s Fable model were lifted last week, and customer access was restored one day later. The lifting of those restrictions ended a period of uncertainty that had limited access for users around the world. CNBC also reported that strict US controls on advanced AI could unintentionally help Chinese AI companies grow faster.
Chinese companies are using the opportunity to offer AI models that are cheaper and easier for businesses to access. One example is Zhipu (Knowledge Atlas Technology JSC), which launched its GLM 5.2 AI model last month. GLM 5.2 is free to download, can be fine-tuned by companies, and can run on an organisation’s own servers, making it an attractive option for businesses.
Overall, the approval of GPT-5.6 shows that the US government is now reviewing powerful AI systems much more closely before allowing companies to release them to the public.
