WASHINGTON: OpenAI said its latest generation of artificial intelligence models, the GPT-5.6 series, will be released to the public on Thursday after a period of limited access, amid heightened scrutiny over the technology's potential cybersecurity risks.
The company announced in a post on X that its new flagship model, GPT-5.6 Sol, along with the mid-range Terra and lower-cost Luna models, will become publicly available while preview access is expanded globally.
The announcement comes after OpenAI said in late June that it had provided preview access to GPT-5.6 to a select group of trusted U.S.-based partners at the request of the U.S. government because of concerns about the models' advanced ability to identify software vulnerabilities that could potentially be exploited by hackers.
According to Axios, citing a source familiar with the matter, the Trump administration approved a broader public launch following technical testing and discussions between OpenAI executives and government officials. OpenAI, the White House and the US Department of Commerce did not immediately comment on the report.
Large language models power chatbots and a growing range of AI applications by processing vast amounts of digital information to generate human-like responses.
The GPT-5.6 lineup includes Sol, OpenAI's most capable model, Terra, designed for general-purpose tasks, and Luna, a faster, lower-cost option intended for high-volume use.
The rollout follows a similar move by rival Anthropic, which last week said it was restoring access to its most advanced AI models, Fable 5 and Mythos 5, after U.S. authorities lifted restrictions on where the systems could be released.
The US government is developing new criteria to determine which advanced AI models should be subject to national security restrictions under a White House executive order, reflecting growing concerns over the potential misuse of increasingly powerful AI systems.
OpenAI has said it does not believe government preview requirements should become the long-term standard, arguing they delay access to advanced AI tools for businesses, developers and other users. The company added that it is working with U.S. officials to establish a repeatable framework for future model releases.
As competition in the AI sector intensifies, OpenAI said Terra will be priced at about half the cost of its predecessor, GPT-5.5, in an effort to attract more customers.
More read, OpenAI signs AI deal with Samsung Korea for ChatGPT & Codex rollout
OpenAI and Anthropic have both confidentially filed paperwork with U.S. regulators for initial public offerings, with reports indicating the companies are targeting valuations approaching $1 trillion, underscoring the growing commercial stakes in the global AI race.
