OpenAI’s newest model, GPT 5.6, won’t get a wide public release. Instead, the company plans to share it only with select partners after the Trump administration pushed for a limited rollout, according to The Information.
CEO Sam Altman told staff that the government would be “approving access customer by customer” during a preview period. If the restricted release goes well, OpenAI hopes to follow with a broader launch a couple of weeks later.
The Office of the National Cyber Director and the Office of Science and Technology Policy were the agencies that reportedly requested the limited release. Their staffers “worked closely” with OpenAI on the rollout plan.
This marks a notable shift. The Trump administration initially positioned itself as taking a “hands-off” approach to AI. But recent months have seen a push for federal oversight. Earlier in June, Trump signed an executive order directing certain AI companies to voluntarily submit new models to the government for testing before public release.
The move mirrors what Anthropic already did with its frontier cyber model Claude Mythos, released only to a small group of partners through a program called Project Glasswing. Whether that’s genuine safety concern or clever marketing is still debated — probably some of both.
The core concern driving the pressure: frontier AI models capable of finding and exploiting software vulnerabilities at machine speed. LLMs have already proven effective at writing malware and, in some cases, executing ransomware attacks autonomously. Models that can do this faster than any human analyst pose an obvious problem for organizations running complex software.
Since these models remain closed to the public, it’s hard to gauge the real threat level. But the government clearly isn’t willing to leave it to chance.
