OpenAI confirmed on Friday that it is delaying the public release of its next-generation AI model, GPT-5.6, at the request of the Trump administration. The company said it will first share the models with a small set of preapproved customers, which will be vetted by the US government. It will then work with the administration to gradually expand access. OpenAI expressed dissatisfaction with the delay but believes it is temporary, aiming to make GPT-5.6 available to the public in the coming weeks.
The Trump administration’s request for OpenAI to stagger the release of its AI models follows an executive order signed earlier this month to address cybersecurity concerns related to powerful AI models. The order called for a voluntary process for AI labs to share models with the government 30 days before broader release. However, OpenAI executives stated that no such framework exists yet, leaving frontier AI labs in an uncertain interim period. The White House is seeking to collaborate with these labs to address cybersecurity risks while balancing regulatory concerns.
OpenAI plans to broaden the list of customers it can share GPT-5.6 with next week, including some international partners. The company emphasized that it cannot disclose how the White House approves these customers, only that it sends a list and receives feedback. GPT-5.6 will be released in three versions: Sol, the most capable; Terra, a middle-tier option; and Luna, a fast and affordable version. The company claims GPT-5.6 Sol is its most capable model yet on cybersecurity, biology, and agentic abilities, supported by a 'layered safeguard stack' to prevent malicious use.
Source: wired
