OpenAI will publicly launch its latest Artificial Intelligence (AI) model, GPT-5.6, on Thursday following a one-month delay requested by the United States (US) government due to national security concerns regarding the potential misuse of powerful AI systems.
The development follows the US government lifting restrictions on Anthropic’s Fable and Mythos AI models last week. Anthropic had previously been ordered to suspend access to the models for a three-week period over national security risks.
Washington has scaled up its scrutiny of advanced AI model releases to identify potential threats, specifically focusing on concerns that the technology could be utilized by military or intelligence operations in China, Russia, and other countries of concern.
According to a report by Axios citing a source familiar with the matter, the US Department of Commerce approved the broad launch of GPT-5.6.
The approval followed additional government testing conducted under Washington’s new oversight framework for frontier AI. Prior to this approval, OpenAI restricted the model’s access to a small group of vetted partners, with details of those partners shared directly with US authorities.
OpenAI announced in a post on X late on Tuesday that it plans to launch three models: GPT-5.6 Sol, Terra, and Luna.
Sol represents OpenAI’s most advanced model to date, while Terra is a mid-tier, lower-cost model, and Luna serves as the most cost-efficient option.
The current framework of increased scrutiny for AI models began after US President Donald Trump signed an executive order.
The order established a voluntary framework requiring AI developers to offer “covered frontier models” to the US government for up to 30 days before releasing them to trusted partners.
Regarding safety protocols, Anthropic has stated that it was “probably impossible” to make any AI model fully robust to jailbreaks, noting the potential for the development of a universal jailbreak that would be able to unblock “an entire class of harmful behaviors”.
