OpenAI GPT-5.6 Cleared for Public Release After US Government Testing and Review

OpenAI has received US government clearance to broadly release GPT-5.6 after weeks of testing by the Commerce Department's Center for AI Standards and Innovation. The model had been restricted to approximately 20 government-approved partners—the first time an American frontier model was gated behind a state-approved roster.

Three-Tier Model Family

GPT-5.6 consists of three distinct models:

• Sol: The flagship model, described by OpenAI as strong at coding, biology, and cybersecurity, with a "max reasoning effort" mode for complex problems
• Terra: A lower-cost mid-tier option aimed at everyday enterprise workloads
• Luna: The fastest and cheapest variant, built for high-volume tasks requiring speed

Pricing details for the three tiers have not been disclosed.

Government Review Process

The review sits within a framework established by the Trump administration on June 2, which introduced a voluntary pre-release check for the most capable models. However, GPT-5.6's review went beyond voluntary—OpenAI agreed to a government-managed access list after being asked to slow the launch.

OpenAI sent technical experts to Washington to answer questions from the Commerce Department, according to Axios. The government's particular focus centered on Sol's capabilities in biology and cybersecurity, areas considered sensitive for national security.

OpenAI has stated it does not believe this government access process should become the long-term default, while agreeing to participate this time.

Commercial and Competitive Implications

The weeks-long delay had real commercial stakes. Every week GPT-5.6 remained in restricted preview was time competitors could court enterprise customers OpenAI wanted to reach with its new tiered offerings.

OpenAI expects to widen access within days and make all three tiers generally available in the coming weeks, though no specific public date has been set. The release builds on the foundation laid by GPT-5.5 earlier this year.

Broader Industry Impact

The same oversight framework covers OpenAI's rivals, meaning this rollout pattern will likely influence how future frontier models from any US company reach the public. The administration has already exercised its power to halt releases, having ordered Anthropic to shut down two models.

What This Means

This marks the first time a leading US AI lab has released a frontier model on the government's schedule rather than its own. The precedent raises questions about whether this was a one-time arrangement or the new normal for frontier AI releases in the United States. For companies racing to maintain competitive advantages in AI, the prospect of government-controlled release timelines introduces a significant new variable in product planning and market strategy. The three-tier approach to GPT-5.6 also signals OpenAI's push to capture different segments of the enterprise market simultaneously, trading raw capability for cost and speed depending on use case.