OpenAI CEO Sam Altman speaks to reporters after meeting with U.S. House Minority Leader Hakeem Jeffries on Capitol Hill in Washington, DC, June 3, 2026.
Brendan Smelowski | AFP | Getty Images
OpenAI said it will publicly release its GPT-5.6 Sol, Terra and Luna models on Thursday, about two weeks after the artificial intelligence company limited deployment to a “small group of trusted partners” at the request of the U.S. government.
“Happy building,” OpenAI CEO Sam Altman wrote in a post on X late Tuesday.
On Wednesday, OpenAI also announced a new generation of voice models called GPT-Live. The company said models can listen and talk at the same time, which means communicating with models “feels much more like a real conversation,” according to the blog post.
Two versions of the voice models, GPT‑Live‑1 and GPT‑Live‑1 mini, will be available to ChatGPT users around the world on Wednesday, OpenAI said.
OpenAI unveiled a series of GPT-5.6 models in June and initially agreed to share them with a select group of organizations whose “participation was shared with the government,” according to a blog post. The company said it believes in “wide access” and will work to make the models more widely available in the coming weeks.
“We do not believe that this government access process should become a long-term standard,” OpenAI said at the time. “It keeps the best tools available to the users, developers, enterprises, cyber defenders and global partners who need them.”
The public launch of GPT-5.6 models comes after OpenAI’s main competitor, Anthropic, restored access to its Claude Fable 5 and Mythos 5 models after a week-long conflict with the government. anthropic had to cut off access to the models to comply with an export control directive that the US Commerce Department lifted late last month.
The Trump administration has taken a more active role in the deployment of artificial intelligence since President Donald Trump signed an executive order on artificial intelligence in June. The order requires AI model developers to voluntarily submit advanced models to the government to evaluate their capabilities before full release, and gives federal agencies 60 days to develop an evaluation process.
In June, OpenAI said it was working with the government to help create a framework for such assessments and develop a “repeatable process for future model releases.”
The company said GPT-5.6 Sol is its “strongest model,” according to a blog post, and that it is more effective in coding, biology and cybersecurity.
OpenAI said in a post on X late Tuesday that it is expanding preview access to GPT-5.6 models worldwide.
LOOK: Tusk Ventures CEO on OpenAI’s 5% stake offer to government: ‘It doesn’t make sense’
Choose CNBC as your preferred source on Google and never miss a beat from the most trusted business news source.