The White House is asking OpenAI to slow down the release of its new model due to safety concerns

OpenAI’s release of its latest model, GPT 5.6, will reportedly be nothing like its previous releases. Instead of distributing it to the public, the company plans to share it only with a select group of close partners because the Trump administration has asked it to, Information reports.

At a meeting this week, CEO Sam Altman reportedly told staff that the government “will approve customer access for every customer” during the preview period. Altman reportedly added that if the limited release goes well, OpenAI hopes to follow it up with a broader public release “a few weeks later.”

In other words, the Trump administration appears to be putting pressure on OpenAI to do what Anthropic is already voluntarily doing: keep its most powerful AI models secret.

According to the information, the new OpenAI model is not only under review by the administration, but its employees have also “worked closely” with the government on the upcoming release. The agencies that reportedly requested a limited release are the Office of the National Cyber ​​Director and the Office of Science and Technology Policy.

The Trump administration — which originally presented itself as taking a “hands-off” approach to AI — has in recent months pushed for federal oversight of new models. Earlier this month, Trump He signed an executive order Directing certain AI companies to voluntarily submit new models to the government for testing and evaluation before releasing them publicly.

Earlier this year, Anthropic generated a fair amount of controversy when it announced that its new frontier cyber model, Cloud Mythos, would be They will just be released to a small coterie of partners through a program called Project Glasswing. Anthropic argued that its model was simply too powerful and could cause more harm than good if it fell into the wrong hands. Observers have since debated whether the Anthropic Rhetoric was merely a marketing gimmick or a legitimate attempt to prevent abuse of a powerful model. The answer may be somewhere in between.

Cybercriminals used automated tools to A very long timeBut in the age of generative AI, they now have more digital repertoire than ever before. LLMs have proven skill in Writing malwareAnd some can even Executing entire ransomware attacks Independently.

The specific concern about leading cyber tools like Mythos is that they are ostensibly able to identify and exploit software vulnerabilities at speeds that no human analyst can match. Since many software systems contain hidden bugs that act as entry points into enterprise networks, this is clearly a significant problem for any organization running complex software infrastructure. However, since these models remain closed to the public, it is difficult to know how much of a threat they actually pose.