The release of OpenAI’s latest model, GPT 5.6, will reportedly not be the same as previous releases. Instead of distributing it to the public, the plan is to share it only with a select group of close partners at the direction of the Trump administration, The Information reported.
At a meeting this week, CEO Sam Altman reportedly told employees that during the preview period, the government would “approve access on a customer-by-customer basis.” Altman added that if the limited release goes well, OpenAI hopes to follow with a general, widespread release “in a few weeks.”
In other words, the Trump administration appears to be pressuring OpenAI to do what Anthropic is already doing voluntarily: keep its most powerful AI models secret.
Not only is OpenAI’s new model under review by the administration, according to The Information, but its staff has “worked closely” with the government on the upcoming launch. Agencies that requested limited disclosure include the Office of the National Cyber Chief and the Office of Science and Technology Policy.
The Trump administration, which originally positioned itself as taking a “hands-off” approach to AI, has been pushing for federal oversight of new models in recent months. Earlier this month, Trump signed an executive order directing certain AI companies to voluntarily submit new models to the government for testing and evaluation before making them public.
Earlier this year, Anthropic sparked quite a bit of controversy when it announced that its new pioneering cyber model, the Claude Mythos, would be released to a small number of partners through a program called Project Glasswing. Anthropic argued that the model was so powerful that in the wrong hands it could do more harm than good. Observers have since debated whether Anthropic’s rhetoric was simply a marketing gimmick or a legitimate attempt to prevent its powerful model from being misused. The answer may lie somewhere in between.
Cybercriminals have long used automated tools, but in the age of generative AI, they now have more digital weapons than ever before. LLMs have proven to be adept at writing malware, and some are even capable of executing entire ransomware attacks autonomously.
A particular concern with cutting-edge cyber tools like Mythos is that they can ostensibly identify and exploit software vulnerabilities at a speed that human analysts can’t match. Many software systems contain hidden bugs that serve as entry points into corporate networks, posing obvious and serious challenges to any organization running a complex software infrastructure. However, because these models are not available to the public, it is difficult to say how much of a threat they actually pose.
If you purchase through links in our articles, we may receive a small commission. This does not affect our editorial independence.
