OpenAI defers public rollout of GPT-5.6 as US seeks early access

OpenAI builds the world’s most capable AI. The US government would like to see it before everyone else does.

On June 25, 2026, CEO Sam Altman disclosed that OpenAI is postponing the broad public release of GPT-5.6, its newest and most advanced model family, at the request of the Trump administration. The delay is rooted in safety and national security concerns, specifically around the cybersecurity capabilities that a model this powerful could put into the hands of bad actors.

Instead of a standard product launch, GPT-5.6 is entering the world through a narrow door. A limited preview will be made available only to a select group of trusted enterprise partners, with the US government approving customer access on a case-by-case basis during this preliminary phase.

What GPT-5.6 actually is

The GPT-5.6 family is not a single model. It ships as three distinct variants built for different use cases.

Sol is the flagship, described by OpenAI as its most capable model to date. Terra is positioned as the cost-efficient option, priced at roughly half the rate of Sol. Luna rounds out the family as the fast, affordable tier for high-volume, latency-sensitive applications.

OpenAI published a system card alongside the limited preview announcement on June 26, 2026, outlining Sol’s performance profile. The model shows significant improvements in agentic coding, biology-related tasks, and cybersecurity applications.

OpenAI says GPT-5.6 ships with an upgraded safety framework.

Why the government is involved at all

This is not the first time Washington has inserted itself into an AI product cycle. OpenAI’s move mirrors pressure applied to Anthropic, where similar restrictions on advanced model access were enacted over comparable security concerns.

Altman’s framing during an internal Q&A was collaborative rather than combative. He described the arrangement as a phased release plan and positioned government engagement as a mechanism for safety feedback.

The limited preview phase is expected to last days to weeks, contingent on government approval timelines. Broader availability through ChatGPT, Codex, and the API is anticipated in the coming weeks, assuming regulatory processes move at a reasonable pace.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.