OpenAI plans to introduce its most advanced artificial intelligence model, GPT-5.6, to the public later this week after a recent delay prompted by federal security concerns. The move comes after the U.S. Department of Commerce granted approval following a period of restricted access involving only government-approved organizations.
OpenAI announced via social media that it would release not only GPT-5.6 Sol—its most sophisticated model to date—but also Terra and Luna, catering to users seeking mid-range and cost-effective AI solutions, respectively.
The upcoming rollout marks the conclusion of an initial phase where use of GPT-5.6 was confined to vetted partners as officials evaluated its potential risks. According to Axios’ report citing an individual familiar with the launch schedule, OpenAI intends to make the model broadly available later in the week, following months of safety discussions and technical reviews with federal agencies.
The U.S. government has been heightening its oversight of advanced AI technologies, citing worries that these systems could be exploited by foreign military or intelligence interests. Concerns have centered on countries such as China and Russia.
The Trump administration last month ordered OpenAI to postpone the release and temporarily restrict access to trusted entities, pending security evaluations. OpenAI stated then that this staggered launch approach was the result of ongoing efforts to establish clear regulatory standards for releasing powerful AI systems.
Federal scrutiny of advanced models has intensified since President Donald Trump issued an executive order, establishing a framework for voluntary disclosure that allows officials up to 30 days to review qualifying AI models before they become widely available.
OpenAI’s experience is not unique, as competitors face similar barriers. According to Axios’ source, rival firm Anthropic also encountered government-mandated delays with the deployment of its Mythos and Fable systems. These measures indicate that individualized reviews are increasingly becoming standard procedure for advanced AI launches while comprehensive regulatory guidelines are still being formulated.
