OpenAI announced Friday that it is holding off on a full public release of its latest artificial intelligence model, GPT-5.6, after the U.S. government requested early access to the technology before it becomes widely available.
Rather than launching to the general public, the company is initially making the model accessible only to a limited group of vetted partners, whose identities have been shared with federal authorities.
The move reflects increasing concern in Washington about the national security implications of powerful AI systems. Policymakers have been pushing tech companies to put stronger safeguards in place around their most advanced tools.
By gaining early access to these so-called frontier AI models, U.S. officials hope to spot potential threats — from cyberattacks to military misuse — before the technology is broadly deployed.
In a blog post, OpenAI described the restricted release as a short-term measure while the company collaborates with the federal government on a wider framework for how future AI models will be launched. The company also noted that it briefed the government on its plans and the model’s capabilities ahead of the release.
President Donald Trump signed an executive order earlier this month creating a voluntary framework that allows AI developers to give the U.S. government access to advanced “covered frontier models” for up to 30 days before sharing them with trusted partners.
“We are taking this short-term step because we believe it is the strongest path to broader availability in the coming weeks, while we work with the Administration to develop the cyber Executive Order framework and a repeatable process for future model releases,” OpenAI stated.
The company said it will continue thorough testing and close coordination with partners as it prepares for a broader release. However, OpenAI also warned that this level of government oversight should not become a permanent standard, and it did not disclose the names of its partner organizations.
OpenAI expressed concern that ongoing government restrictions could limit access to advanced AI tools for developers, businesses, cybersecurity professionals, and international partners who stand to benefit from them.
The centerpiece of the new model lineup is GPT-5.6 Sol, described as the company’s most capable model to date. It is accompanied by a mid-range option called Terra and a lower-cost version called Luna.
