Why is OpenAI delaying the broad GPT-5.6 launch?

OpenAI delayed the broad GPT-5.6 launch at the request of the Office of the National Cyber Director and OSTP under the Trump administration. Officials asked OpenAI to restrict early access while they assess national-security risks of the frontier model. OpenAI says the delay is temporary and broader availability is planned within weeks.

Who can use GPT-5.6 during the limited preview?

GPT-5.6 access during the limited preview is restricted to about 20 partners whose identities OpenAI shared with the US government. OpenAI has not published the list. The Trump administration will approve additional customers one by one before any broader rollout, according to internal remarks Sam Altman gave staff.

How does this differ from the standard OpenAI preview?

GPT-5.6 is the first OpenAI model run through the voluntary frontier-AI framework set up by a May 2026 executive order. Earlier previews were limited at OpenAI's discretion. With GPT-5.6, the government gets up to 30 days of early access before any trusted partner can use the model, and must sign off on each new customer.

Is this related to the Claude Mythos 5 export controls?

Yes, both moves are part of the same Trump-administration framework for frontier AI. Two weeks ago Commerce blocked Anthropic's Claude Mythos 5 and is now restoring it only for about 100 Annex A partners. GPT-5.6 follows the same shape: federal sign-off required before any wider release of a top-end US frontier model.

OpenAI · 2026-06-26 · major

GPT-5.6 rollout delayed — US government will vet every customer

OpenAI postponed the broad GPT-5.6 launch at the Trump administration's request, limiting initial access to about 20 government-vetted partners. The Office of the National Cyber Director will approve customers one by one.

OpenAI's GPT-5.6 ships only to government-approved partners while a Trump-era frontier AI framework decides who gets in.

Quick facts

Model	GPT-5.6 (Sol, Terra, Luna)
Maker	OpenAI
Initial access	About 20 government-vetted partners
US offices involved	Office of the National Cyber Director, OSTP
Framework	Voluntary 30-day pre-release review (May 2026 executive order)
Approval cadence	Customer by customer
Announced	June 26, 2026

What is it?

OpenAI's GPT-5.6 preview is gated behind a new US government vetting step. The company has paused a broad public launch at the request of the Trump administration's Office of the National Cyber Director and Office of Science and Technology Policy. Access is limited to roughly 20 vetted partners whose identities were shared with federal officials.

How does it work?

A May 2026 executive order set up a voluntary framework: covered frontier-model developers can offer their newest models to the US government for up to 30 days before any trusted-partner release. GPT-5.6 is the first model OpenAI is running through that process. Sam Altman told staff the government will approve access customer by customer rather than signing off on a whole tier at once.

Why does it matter?

GPT-5.6 was already a limited preview, but the delay shows the federal government — not OpenAI — now decides who gets early access to the strongest US frontier models. The vetting mirrors the Annex A regime around Anthropic's Claude Mythos 5, signaling that Washington-led pre-release review is becoming the default for any new flagship from a top US lab.

Who is it for?

Enterprise GPT-5.6 prospects, policy teams, AI compliance leads

Frequently asked questions

Why is OpenAI delaying the broad GPT-5.6 launch?: OpenAI delayed the broad GPT-5.6 launch at the request of the Office of the National Cyber Director and OSTP under the Trump administration. Officials asked OpenAI to restrict early access while they assess national-security risks of the frontier model. OpenAI says the delay is temporary and broader availability is planned within weeks.
Who can use GPT-5.6 during the limited preview?: GPT-5.6 access during the limited preview is restricted to about 20 partners whose identities OpenAI shared with the US government. OpenAI has not published the list. The Trump administration will approve additional customers one by one before any broader rollout, according to internal remarks Sam Altman gave staff.
How does this differ from the standard OpenAI preview?: GPT-5.6 is the first OpenAI model run through the voluntary frontier-AI framework set up by a May 2026 executive order. Earlier previews were limited at OpenAI's discretion. With GPT-5.6, the government gets up to 30 days of early access before any trusted partner can use the model, and must sign off on each new customer.
Is this related to the Claude Mythos 5 export controls?: Yes, both moves are part of the same Trump-administration framework for frontier AI. Two weeks ago Commerce blocked Anthropic's Claude Mythos 5 and is now restoring it only for about 100 Annex A partners. GPT-5.6 follows the same shape: federal sign-off required before any wider release of a top-end US frontier model.

Try it

GPT-5.6 access only via OpenAI's vetted-partner application