OpenAI Ceo Sam Altman speaks to journalists after assembly with US Home Minority Chief Hakeem Jeffries on Capitol Hill in Washington, DC, on June 3, 2026.
Brendan Smialowski | AFP | Getty Photographs
OpenAI stated it would publicly launch its GPT-5.6 Sol, Terra and Luna fashions on Thursday, roughly two weeks after the unreal intelligence firm restricted the rollout to a “small group of trusted companions” on the request of the U.S. authorities.
“Completely satisfied constructing,” OpenAI CEO Sam Altman wrote in a publish on X late Tuesday.
OpenAI introduced the fashions in June, and it initially agreed to launch them to a choose group of organizations whose “participation has been shared with the federal government,” in response to a weblog publish. The corporate stated it believes in “broad entry,” and that it could work to make the fashions extra extensively obtainable over the approaching weeks.
“We do not imagine this type of authorities entry course of ought to develop into the long-term default,” OpenAI stated on the time. “It retains one of the best instruments from customers, builders, enterprises, cyber defenders, and world companions who want them.”
The general public launch of the GPT-5.6 fashions comes after OpenAI’s chief rival, Anthropic, restored entry to its Claude Fable 5 and Mythos 5 fashions following a weeks-long conflict with the federal government. Anthropic needed to disable entry to the fashions to adjust to an export management directive, which the U.S. Division of Commerce lifted late final month.
The Trump administration has taken a extra energetic position in AI deployments since President Donald Trump signed an AI govt order in June. The order asks AI mannequin builders to voluntarily present cutting-edge fashions to the federal government to evaluate their capabilities forward of a full launch, and it gave federal businesses 60 days to develop an analysis course of.
OpenAI stated in June that it is working with the federal government to assist set up a framework for such assessments and to develop a “repeatable course of for future mannequin releases.”
The corporate stated that GPT-5.6 Sol is its “strongest mannequin but,” in response to a weblog publish, and that it’s extra succesful throughout coding, biology and cybersecurity.
OpenAI stated in a publish on X late Tuesday that it’s increasing preview entry to the GPT-5.6 fashions globally.
WATCH: Tusk Ventures CEO on OpenAI proposing a 5% stake to the federal government: ‘It makes zero sense’
