OpenAI to unveil GPT-5.6 on July 9 after US security concerns delay

ChatGPT-maker OpenAI is set to release its most powerful Artificial Intelligence (AI) model, the GPT-5.6, on July 9 after the US government requested a delay over national security concerns and potential misuse. In addition, the company has also confirmed to launch GPT-5.6 Sol, and the lower-cost Terra and Luna models.

The public rollout was announced about two weeks after the company was advised of a limited release to a “small group of trusted partners,” as advised by the Trump administration. It is further reported that OpenAI conducted additional testing and held meetings with U.S. government officials to discuss the model and any security concerns. 

Must read: OpenAI proposes giving Trump Administration a 5% stake

OpenAI GPT 5.6 release

When OpenAI first previewed GPT-5.6, it revealed that the model offers improved agentic capabilities and that it performs well in three areas: coding, biology, and cybersecurity. 

On the other hand, the company further said that GPT-5.6 Sol is a specialised version similar to Anthropic's Mythos Preview on ExploitBench, a benchmark used to evaluate how well AI models can understand and solve cybersecurity exploitation tasks.

Now, the company has revealed that these models will be released to the public on Thursday, July 9, 2026, after brief security concerns. 

Must read: Anthropic, OpenAI customers overbilled by $1.7 million due to billing errors

Its rival Anthropic also faced similar rollout restrictions with Claude Fable 5 and Mythos 5. Now, both models have been rolled out to the relevant users, making powerful AI models broadly available.

US national security concerns

After Anthropic’s Mythos Preview, the US government has been more concerned about cybersecurity and misuse of powerful AI models. In June, President Donald Trump signed an AI executive order that involved overseeing the release of advanced AI models. The order requested AI companies to voluntarily share their newest, most powerful AI models with the government before they are released publicly.

The order also gave federal agencies 60 days to conduct a formal evaluation of these AI models. OpenAI agreed to cooperate with the U.S. government to help design this evaluation framework and create a standard that can be used before future AI model launches.