OpenAI will launch its GPT-5.6 model to the general public this Thursday. The authorization comes from the Trump administration following a period of restricted access and government-mandated testing, according to a source familiar with the matter reported by Axios.
The release consists of three tiers: the flagship “Sol” model and two lower-tier versions, “Terra” and “Luna.”
The Washington On-Site Audit
The path to public availability required a series of technical evaluations conducted by the Center for AI Standards and Innovation within the Department of Commerce. To clear the final hurdles, OpenAI deployed technical experts to Washington, D.C. These engineers remained on-site to address potential questions before restrictions were lifted.
It is a reversal of a strategy pushed by the administration last month. At that time, the government pressured OpenAI into a staggered rollout, limiting initial access to government-approved entities. OpenAI stated then that this phased approach was not the company’s preferred method for releasing new models.
Governance by Negotiation
The “green light” for the Sol, Terra, and Luna tiers was contingent on additional testing and meetings between the company and government officials. The Center for AI Standards and Innovation acted as the auditing body.
OpenAI noted that both the company and the government are currently operating before more concrete standards for releasing such models—called for in President Trump's latest AI executive order—have been finalized.
AI as a National Security Asset
GPT-5.6 is the latest model to face government intervention. Anthropic has encountered similar friction.
In June, the Commerce Department banned foreigners from having access to Anthropic’s “Mythos” and “Fable” models, essentially forcing their withdrawal from the market. Although the ban on Fable was lifted last week—with customer access restored one day later—the episode highlights a trend of treating AI capabilities as sensitive national security assets.
The Price of Market Access
The presence of OpenAI engineers in D.C. during the GPT-5.6 testing phase underscores this level of direct oversight.
OpenAI has confirmed the suite will be available this Thursday. Details on the specific capabilities of Sol, Terra, and Luna are expected upon the official release.
Official announcements can be monitored via the U.S. Department of Commerce or OpenAI’s newsroom.