WASHINGTON, July 8 (DC Times Online) — OpenAI is letting only a small group of trusted partners use its newest AI model for now, after the Trump administration asked the company to slow the launch while officials review possible cybersecurity risks.
For most users, that means the model is not yet a public product. OpenAI says it expects broader access in the coming weeks if the preview goes smoothly.
What is OpenAI releasing?
OpenAI said the limited preview covers its newest model in the GPT-5.6 line, which it identified in public statements as GPT-5.6 Sol. The company described it as its strongest model yet.
But OpenAI also said the model is being released carefully because more capable AI systems can create new security concerns. In its statement, the company said the model is better at helping people find and fix software vulnerabilities than at carrying out cyberattacks. OpenAI said the model does not cross its internal “Cyber Critical” threshold, a company test for especially risky cyber behavior.
Why is the White House involved?
The Trump administration’s review is tied to a June executive order signed by President Donald Trump. The order created a voluntary federal framework for reviewing advanced AI models before they are widely released.
The White House said the process is meant to help officials look for risks to financial systems, national security, and other critical systems. Reporting on the request identified the Office of the National Cyber Director and the Office of Science and Technology Policy as the two White House offices involved.
What does the limited rollout mean in practice?
The short answer: fewer people can test the model, and the public cannot use it yet.
OpenAI said the first users are a small group of trusted partners. The company said those partners’ participation has been shared with the government. OpenAI has not publicly named the partners.
That means the model is still in a preview stage. It is not a full release, and it is not being rolled out broadly through ChatGPT or to the general public right now.
Why would a company do this?
A limited rollout lets OpenAI and federal officials watch how the model behaves before making it more widely available. The main concern is that a powerful model could help defenders find flaws in software, but could also be used to probe systems more aggressively.
OpenAI said it does not want this kind of government access process to become the long-term default. Still, the company said the preview is the strongest path to broader availability in the coming weeks.
How unusual is this?
It is a sign that frontier AI models are increasingly being treated like sensitive technology, not just consumer software. The government is not banning the model. Instead, it is asking for a slower, more controlled release while it reviews the risks.
That puts the model in a middle category: available, but only to a small set of approved users for now.
For anyone waiting to try it, the key point is simple. OpenAI says the model is coming, but not all at once. First comes a limited preview. Broader access may follow after the review period ends.
