Who approved this limited release approach?

The Office of the National Cyber Director and the Office of Science and Technology Policy requested the limited release. OpenAI's staff also worked closely with the government on the upcoming release plan.

When will the broader public release happen?

OpenAI hopes to follow with a general, broader release a "couple of weeks later" if the limited release preview goes well.

Is OpenAI the only company doing this?

No. Anthropic is already voluntarily keeping its most powerful AI models under wraps through a program called Project Glasswing, which limits access to its frontier cyber model Claude Mythos to a small group of partners.

Large Language ModelsAI Business & IndustryTechCrunch AIJun 26, 2026Jun 26, 2026, 10:00 JST1 min read

The Trump administration is pressuring OpenAI to limit early access to its new GPT 5.6 model to select partners before a broader public release, shifting from its previous hands-off AI stance.

Send on LINE

Key takeaway

OpenAI is being required by the Trump administration to limit early access to its new GPT 5.6 model to approved partners before a public release.
This represents a reversal of the administration's previous hands-off stance and reflects growing federal pressure on AI companies to submit models for government testing before public deployment.

3 Key Points

What happened
OpenAI CEO Sam Altman told staff that the Trump administration will be "approving access customer by customer" during a preview period for the company's new GPT 5.6 model. The Office of the National Cyber Director and the Office of Science and Technology Policy requested the limited release, and staff worked closely with the government on the upcoming release plan.
Why it matters
The Trump administration, which originally positioned itself as taking a "hands off" approach to AI, has in recent months pushed for federal oversight of new models. This shift aligns with an executive order Trump signed directing certain AI companies to voluntarily submit new models to the government for testing and evaluation before releasing them publicly.
What to watch
OpenAI hopes to follow the limited release with a broader, general release a "couple of weeks later" if the preview period goes well. This mirrors Anthropic's approach with its frontier cyber model Claude Mythos, which has been released only to a small group of partners through Project Glasswing.

Ask the AI about this article →

Summaries like this, in your inbox every morning.

Get news like this every morning — free

FAQ

Who approved this limited release approach?: The Office of the National Cyber Director and the Office of Science and Technology Policy requested the limited release. OpenAI's staff also worked closely with the government on the upcoming release plan.
When will the broader public release happen?: OpenAI hopes to follow with a general, broader release a "couple of weeks later" if the limited release preview goes well.
Is OpenAI the only company doing this?: No. Anthropic is already voluntarily keeping its most powerful AI models under wraps through a program called Project Glasswing, which limits access to its frontier cyber model Claude Mythos to a small group of partners.

TechCrunch AIRead Original Article