Background and Rollout Sequence
OpenAI on July 9, 2026 began a broad public rollout of GPT-5.6, its newest Large Language Model (LLM) generation, after an approximately 13-day restricted preview coordinated with the U.S. government. <cite index="9-8,9-9">As part of its ongoing engagement with the U.S. government, OpenAI previewed its plans and the models' capabilities ahead of launch, and at the government's request started with a limited preview for a small group of trusted partners whose participation was shared with federal officials.</cite>
<cite index="13-1,13-2">The U.S. Department of Commerce cleared OpenAI for a broad public rollout of GPT-5.6, ending a period of regulatory uncertainty that had delayed the AI system's wider release.</cite> <cite index="16-9">Technical staff from OpenAI worked directly with the Commerce Department's Center for AI Standards and Innovation during the review, responding to questions raised by government officials before approval was granted.</cite>
The review framework has broader policy context. <cite index="16-7">President Donald Trump signed an executive order establishing a voluntary framework that allows developers of certain frontier AI models to provide them to the U.S. government for review before broader deployment.</cite> <cite index="14-14">The voluntary framework allows developers to provide "covered frontier models" to the U.S. government for up to 30 days before releasing them to trusted partners.</cite>
The Three-Tier Model Family
<cite index="3-1">GPT-5.6 spans three model tiers: Sol, the flagship; Terra, a lower-cost model with performance competitive with GPT-5.5; and Luna, the fastest and most affordable model.</cite> <cite index="1-17">In this new naming system, the number identifies a model's generation, while Sol, Terra, and Luna identify durable capability tiers that can advance on their own cadence.</cite>
<cite index="3-21,3-22">GPT-5.6 is available starting July 9 across ChatGPT, Codex, and the OpenAI Application Programming Interface (API), with the rollout beginning globally and continuing gradually toward full availability over 24 hours.</cite> <cite index="1-19">GPT-5.6 is priced per one million tokens: Sol at $5 input / $30 output; Terra at $2.50 input / $15 output; and Luna at $1 input / $6 output.</cite>
Benchmark Performance
<cite index="3-10,3-11">GPT-5.6 Sol is OpenAI's best coding model yet; on the Artificial Analysis Coding Agent Index, Sol with max reasoning sets a new state of the art at 80, 2.8 points above Fable 5, while using less than half the output tokens, taking less than half the time, and costing about one-third less.</cite>
<cite index="3-13">GPT-5.6 Sol sets new state-of-the-art results on BrowseComp at 92.2% and OSWorld 2.0 at 62.6%.</cite> On science tasks, <cite index="1-9">on GeneBench v1, which evaluates long-horizon genomics and quantitative-biology analyses, Sol achieves stronger results than GPT-5.5 while using fewer tokens.</cite>
Safety Architecture and Cybersecurity Classification
<cite index="9-11,9-12">Under OpenAI's Preparedness Framework, Sol, Terra, and Luna are classified as "High" capability in both Cybersecurity and Biological and Chemical risk, though none reach the High threshold in AI Self-Improvement.</cite> <cite index="9-15,9-16">The models represent a meaningful step up in cybersecurity capability but do not reach the risk framework's highest level (Critical); Sol and Terra can find vulnerabilities and pieces of exploits, but in cybersecurity testing were unable to carry out autonomous, end-to-end attacks against hardened targets.</cite>
<cite index="5-17">OpenAI says it conducted its most intensive security testing to date ahead of the launch, including approximately 700,000 Graphics Processing Unit (GPU)-hours of automated red teaming.</cite> <cite index="5-15,5-16">According to the company, Sol's cybersecurity safeguards block about ten times more potentially harmful activity than before, and for the most demanding security tasks, access remains limited to verified users within OpenAI Daybreak's Trusted Access program.</cite>
Policy Implications and OpenAI's Position
OpenAI has publicly distanced itself from the review process becoming standard practice. <cite index="14-10,14-11">The company stated: "We don't believe this kind of government access process should become the long-term default. It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them."</cite>
<cite index="11-10,11-11">There is still no agreement on what kinds of models require government scrutiny, or what agency or agencies should perform those evaluations; for now, the Department of Commerce's Center for AI Standards and Innovation appears to be taking the lead, but a Trump executive order instructs six cabinet agencies to determine a final process by early August.</cite>
<cite index="6-27">OpenAI also announced it is launching GPT-5.6 Sol on Cerebras at up to 750 tokens per second in July, with access initially limited to select customers as capacity expands.</cite>