OpenAI has begun offering a preview version of its new AI model, "GPT-5.6," to a small set of government-approved trusted partners. The staggered rollout follows a request from the Trump administration, with a wider release expected within the coming weeks.
June 25–26, 2026 · OpenAI
OpenAI Staggers GPT-5.6 Release at the White House's Request
OpenAI is offering a limited preview of its next flagship model to government-approved, trusted partners first — approved "customer by customer" — before a wider rollout expected within weeks.
30 days
Voluntary government review window per the June 2026 Executive Order
1.5M
Reported context window in tokens (unconfirmed)
10–15%
Reported token-efficiency gain (unconfirmed)
Context window: GPT-5.5 → GPT-5.6 (reported)
Roughly +50% context, plus reported gains in agentic workflows, multi-hour session reliability, and coding/agent execution.
The staggered-release path
Executive Order
Voluntary pre-release review of frontier models
→
Limited preview
Approved "customer by customer"
→
Wider release
Expected within coming weeks
The stated rationale
National security concerns over the capabilities and risks of cutting-edge models. The request was led by the Office of the National Cyber Director and OSTP.
Open debate
Security vs. innovation, and parallels to export-style controls — echoing Anthropic's earlier restriction of its Mythos and Fable models.
All specs are unconfirmed. No official announcement or model card has been published — benchmarks (SWE-bench, Terminal-Bench, FrontierMath), pricing, and API availability remain undisclosed, and OpenAI's release notes currently contain no mention of GPT-5.6.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…