On Thursday, OpenAI's CEO, Sam Altman, shared insights with CNBC regarding the company’s newest AI model, GPT-5.6 Sol. He highlighted that this latest iteration demonstrates a 54% increase in token efficiency for agentic coding tasks and claims it rivals or surpasses other models available in the market.
Altman noted, "Every enterprise is now reflecting on their AI expenditures and the returns they're receiving; that's our primary focus."
The launch of the latest suite of models—GPT-5.6 Sol, Terra, and Luna—was initiated on Thursday, following a prior announcement made last month. To comply with U.S. government requests, the initial release has been restricted to a select group of trusted partners.
He elaborated that OpenAI collaborated closely with key government officials, including Commerce Secretary Howard Lutnick, Treasury Secretary Scott Bessent, and U.S. National Cyber Director Sean Cairncross, throughout the approval process. This collaboration, according to Altman, involved a "collaborative back and forth," allowing the government to conduct tests and identify concerns for OpenAI to resolve.
"If broad access is our goal—as it is—and we possess powerful models, we must have confidence in our safety assurances; otherwise, the world could quickly become a troubling place," Altman emphasized.
