OpenAI Unveils New AI Models: Reasons You’re Unable to Access Them

OpenAI is postponing the public launch of its upcoming AI models, GPT-5.6, following a request from Trump’s White House, the company announced on Friday. OpenAI stated that it will initially provide the models to a select group of customers, who will be vetted by the US government. It will then collaborate with the administration to gradually broaden access.

According to a source familiar with the situation, OpenAI is not pleased with this arrangement but believes that the delay and the government approval process are temporary. In a blog post, the company expressed hope that GPT-5.6 would be available to the public shortly. The news of OpenAI’s delay due to the Trump administration’s request was first reported by The Information.

“We don’t think this type of government access process should become the standard practice,” OpenAI noted in its blog post. “It restricts the best tools from users, developers, enterprises, cyber defenders, and global partners who need them. We are undertaking this short-term measure because we believe it is the best route to wider availability soon, while we cooperate with the Administration to create the cyber Executive Order framework and a repeatable process for future model releases.”

Earlier this month, President Trump enacted an executive order addressing the cybersecurity risks associated with advanced AI models. The order stipulated that the White House would establish a “voluntary process” allowing AI labs to share their models with the government 30 days prior to a wider release. The mandate included a provision indicating that the US government would not turn its voluntary process into a de facto licensing regime for AI model releases. However, during a Friday briefing, OpenAI executives stated that no such voluntary framework currently exists. Consequently, leading AI labs find themselves in an unusual interim phase where engaging with the US government regarding AI model launches does not appear entirely voluntary.

The White House has requested OpenAI to stagger its AI model releases just two weeks after sending an export control directive to Anthropic, which led the company to take its most advanced AI models offline for all users. Anthropic’s conflict with the White House remains unresolved, leaving some of its employees unable to access its most sophisticated AI models.

The Trump administration’s request for OpenAI and Anthropic to limit access to their most advanced AI models has created an uncertain landscape for other US AI labs. Over the past two years, the Trump administration has aimed to reduce regulations and bureaucratic hurdles that could hinder America’s AI innovation and diminish competitiveness with China. However, in recent months, the White House has become increasingly apprehensive about the cybersecurity capabilities of new AI models and is seeking solutions to these concerns. A White House official informed WIRED that the administration continues to engage with leading AI labs.

OpenAI intends to expand its customer base for GPT-5.6 next week, including some international partners. OpenAI executives stated that they cannot disclose details on how the White House is approving these customers—the company merely submits a list to the US government and receives feedback.

OpenAI announces that its GPT-5.6 AI models will have three variations: Sol, its most advanced model; Terra, a mid-tier version; and Luna, a quick and cost-effective option. The company asserts that GPT-5.6 Sol is its most powerful model yet in benchmarks assessing cybersecurity, biology, and agentic capabilities. Alongside these new features, OpenAI has implemented a “layered safeguard stack” designed to prevent malicious entities from utilizing its AI model for cyberattacks and other harmful activities.

Update 06/26/26 2:20 pm ET: This story has been updated to include comments from the White House.