OpenAI is scheduled to publicly release its most advanced artificial intelligence model, GPT-5.6, on Thursday after delaying the launch last month following a request from the US government over concerns about the potential misuse of powerful AI systems.
The launch comes shortly after Washington lifted restrictions on Anthropic’s latest Fable and Mythos AI models, which had faced temporary limits due to national security concerns.
The US government has increased oversight of advanced AI technologies as officials assess risks that such systems could be exploited by military or intelligence agencies in countries including China and Russia.
Axios, which first reported the OpenAI launch, said a source familiar with the matter confirmed that the US Department of Commerce approved the wider release of GPT-5.6 after additional government testing under a new oversight framework for advanced AI models.
Before the approval, OpenAI had provided limited access to the model to a selected group of vetted partners, with details shared with government authorities.
OpenAI announced that it would release GPT-5.6 Sol alongside two additional models, Terra and Luna, in a post on X on Tuesday.
The company described Sol as its most advanced model, while Terra is designed as a mid-range, lower-cost option and Luna as the most affordable version.
The White House and the US Department of Commerce did not immediately respond to Reuters’ requests for comment outside regular working hours.
Increased scrutiny of AI systems followed US President Donald Trump’s executive order creating a voluntary framework that allows developers of “covered frontier models” to submit their systems to the US government for up to 30 days before providing access to trusted partners.
Anthropic has previously warned that making AI models completely resistant to jailbreak attempts may be “probably impossible” and highlighted the risk of a universal jailbreak that could unlock a broad range of harmful capabilities across multiple AI systems.
