The newest model of OpenAI will not be like the previous versions of GPT 5.6 release. Instead of releasing it publicly, the company plans to share it only with a select group of close associates because the Trump administration is telling it: This is reported by The Information.
At a meeting this week, CEO Sam Altman told employees the government would “approve client-to-client access” during a review period. If the limited release goes well, OpenAI hopes to follow up with a general, wider release “in a few weeks,” Altman added.
In other words, the Trump administration is pressuring OpenAI to do what Anthropic already voluntarily does: keep its most powerful AI models private.
According to The Information, OpenAI’s new model is not only being vetted by the administration, but its staff has been “working closely” with the government on the upcoming release. The agencies requesting the restricted release are said to be the Office of the National Cyber Director and the Office of Science and Technology Policy.
The Trump administration, which initially positioned itself as taking a “hands-on” approach to artificial intelligence, has in recent months demanded federal oversight of new models. Earlier this month, Trump signed the order directing certain AI companies to voluntarily submit new models to the government for testing and evaluation before releasing them to the public.
Earlier this year, Anthropic caused no small controversy when it announced the launch of its new frontier cyber model, the Claude Mythos. just be free to a small group of partners through a program called Project Glasswing. Anthropic argued that its model was simply too powerful and could do more harm than good in the wrong hands. Observers have since debated whether Anthropic’s rhetoric was just a marketing ploy or a legitimate attempt to avoid abusing a powerful model. The answer may be in the middle.
Cybercriminals have used automated tools for a very long timebut in the age of generative artificial intelligence, there is now more digital ammunition than ever before. LLMs have proven their skills write malwaresome may even perform all ransomware attacks autonomously.
A particular concern with frontier cyber tools like Mythos is that they have the ability to both identify and exploit software vulnerabilities at a speed that no human analyst can match. As many software systems contain hidden bugs that act as entry points to enterprise networks, this presents an obvious and significant challenge for any organization operating a complex software infrastructure. However, since these models remain closed to the public, it’s hard to say how much of a threat they really are.
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.
