OpenAI has confirmed that it is delaying the public rollout of its latest AI models, GPT-5.6, following a request from the Trump administration. The company stated it would initially share these models with a select group of preapproved customers under government oversight, hoping to widen access in the coming weeks.
Although OpenAI’s leadership is not pleased with the delay, they regard this step as necessary to ensure broader availability in the future. In a subsequent blog post, OpenAI expressed concerns about the government’s approach, emphasizing that such access protocols should not become the standard practice, as they limit access to vital tools for users, developers, and enterprises.
The Trump administration’s request came shortly after the White House signed an executive order focused on addressing the cybersecurity risks associated with advanced AI models, mandating a “voluntary process” for AI labs to report their models 30 days prior to public release. Despite this order, OpenAI indicated that a clear voluntary framework does not yet exist, leading to confusion in the AI sector.
This environment of uncertainty was intensified just weeks before the OpenAI announcement when Anthropic, another AI company, was compelled to take its advanced models offline at the government’s request, creating challenges for other U.S. AI labs as well. Historically, the Trump administration had aimed to foster AI innovation while also recently expressing rising concerns about the cybersecurity implications of new AI technologies.
OpenAI plans to broaden its customer base for GPT-5.6 in the coming week, aiming to include some international partners. The forthcoming AI models will consist of three variations: Sol (the top-tier model), Terra (a mid-range version), and Luna (a cost-effective option), all designed with enhanced capabilities and cybersecurity protections to mitigate potential misuse.
For further details, visit OpenAI’s official announcements and reports on the implications of the new executive order.
