OpenAI's GPT-5.6 family debuts in limited preview for trusted partners

OpenAI announced the limited preview of its newest AI models, GPT-5.6 Sol, Terra, and Luna, at the request of the White House. The company will initially provide access only to a small group of trusted partners whose participation has been shared with the US federal government. In a statement, OpenAI emphasized that this is a short-term measure, not a long-term solution to the Trump administration's concerns about powerful AI models falling into the wrong hands. The company aims to work with the administration to develop a repeatable process and a cyber Executive Order framework, with broader availability expected in the coming weeks.

This announcement comes amid escalating tensions between the White House and AI developers. Despite Trump's rhetoric of a less regulatory, innovation-centric approach, his administration has recently taken heavy-handed measures, including forcing Anthropic to disable its Fable 5 and Mythos 5 models for all foreign nationals. Over 100 AI executives and experts signed a letter expressing concerns about the lack of transparency and democratic accountability. Critics warn that such power grabs threaten the open development of AI technology, even as the US promotes coalitions like Pax Silica to strengthen AI leadership globally.