The U.S. Department of Commerce has cleared OpenAI to release GPT-5.6 to the public, with the broad launch expected on Thursday, July 9, 2026, according to reporting from Axios, CNBC, and The Next Web. The decision ends a stretch of federal review that had kept OpenAI’s newest model bottled up in a limited preview, and it clears the way for one of the most closely watched consumer AI releases of the year to reach a general audience.
The path to Thursday was unusual for a commercial software product. Late in June, the White House asked OpenAI to limit the release rather than push GPT-5.6 out to everyone at once. In response the company shipped the model only as a restricted preview, available by most reporting to roughly 20 government approved companies while officials studied what the system could do. That preview status is what Thursday’s clearance lifts, moving GPT-5.6 from a tightly controlled trial into broad availability. OpenAI has argued that restrictions of this kind should not become the default for frontier models, even as it worked through the process.
The evaluation was run by the Center for AI Standards and Innovation, a body housed inside the Commerce Department that has taken on the job of vetting advanced systems. The review centered on export control, dual use, and national security questions tied to the capabilities of the most powerful models, rather than the everyday consumer features most users will notice first. Reporting from Axios, relayed by CNBC, said OpenAI sent technical experts to Washington to answer the agency’s questions, and that the wider deployment was cleared only after additional testing, consultations with officials, and a look at the model’s safety profile.
What comes out of the preview is a family of three models rather than a single system. GPT-5.6 Sol is the flagship and the most capable of the group, positioned by OpenAI as its strongest model to date, with gains in coding, scientific reasoning, biology workflows, and cybersecurity. GPT-5.6 Terra sits in the middle as a balanced everyday model, and GPT-5.6 Luna is the fastest and lowest cost option, built for tasks where speed and price matter more than raw capability. The lineup is meant to let developers and businesses match the model to the job rather than pay for the top tier every time.
The reason the government took an interest sits in the same capabilities OpenAI is marketing. Under the company’s own Preparedness Framework, Sol, Terra, and Luna are all treated as “High” capability in both Cybersecurity and Biological and Chemical risk, the internal thresholds OpenAI uses to flag models that could meaningfully help with cyberattacks or with the creation of biological or chemical threats. A model strong enough to accelerate legitimate biology and security research is, by the same logic, strong enough to warrant a closer look before it reaches millions of people, which is why the review carried weight rather than being a formality.
The clearance also lands in a crowded week. The same Thursday is expected to bring a public release of Grok 4.5 from xAI, now folded into SpaceXAI, setting up a direct collision between two of the most aggressive labs in the field on a single day. For OpenAI, getting the federal sign off in hand before that moment matters, since a delay would have handed rivals an open field while GPT-5.6 sat in preview. The competitive timing helps explain why the company pressed to move from a restricted trial to broad availability as quickly as the review allowed.
For most users the practical effect is simple. Starting Thursday, GPT-5.6 in its Sol, Terra, and Luna versions is expected to be available broadly rather than to a short list of vetted partners, and the tiered structure gives businesses a way to weigh cost against capability. The larger story is about process. A frontier model built by a private company was held back at the government’s request, evaluated by a federal standards body on national security grounds, and then released only once that body signed off, a sequence that offers a preview of how the next generation of powerful systems may reach the public. Whether that becomes a repeatable model or a one time exception is the question OpenAI and its regulators will keep answering with each release that follows.
Source: U.S. Department of Commerce clearance via FinanceFeeds and CNBC/Axios; limited preview background via TechCrunch and CNN; model lineup and deployment safety details via OpenAI and VentureBeat. Credit: OpenAI / U.S. Department of Commerce.
Your First 10 AI Skills
10 practical AI skills, copy-paste prompts and a 7-day plan to start using AI with confidence.
Download the guide →