OpenAI is opening a limited preview of GPT-5.6, led by Sol, its new flagship model, alongside Terra for lower-cost everyday work and Luna for faster, cheaper workloads. The preview starts with a small group of trusted partners, with access initially through the API and Codex, while broader access for ChatGPT, Codex, and API users is planned in the coming weeks.
— Sam Altman (@sama) June 26, 2026Good new first: Sol is a smart, efficient, and a significant step forward. It is the same price as GPT-5.5. Also launching in the GPT-5.6 family is Terra, with 5.5-level performance at half the price.
Bad news: at the request of the US government, it is launching today in…
GPT-5.6 Sol arrives as OpenAI’s strongest model in the new family, with gains positioned around agentic coding, biology workflows, and cybersecurity tasks. The model adds a new max reasoning effort for deeper problem solving and an ultra mode that uses subagents to work on complex tasks beyond a single-agent setup. OpenAI says Sol sets a new state of the art on Terminal-Bench 2.1 and shows stronger GeneBench v1 results than GPT-5.5 while using fewer tokens.
The release is being handled as a controlled rollout because of the model’s cyber and biological capabilities. OpenAI says Sol, Terra, and Luna are classified as High capability in both Cybersecurity and Biological and Chemical risk under its Preparedness Framework, while not reaching the High threshold for AI self-improvement or the Cyber Critical threshold. In browser exploit tests involving Chromium and Firefox, Sol identified bugs and exploitation primitives but did not autonomously produce a full-chain exploit under the tested conditions.
OpenAI is pairing the model upgrade with a layered safeguard stack covering:
- Model-level refusal behavior
- Real-time cyber and biology misuse classifiers
- Account-level review
- Differentiated access
- Monitoring, enforcement, and ongoing testing
Some preview users may see blocked requests or slower responses when generation is paused for extra review, especially in dual-use security contexts where defensive and offensive work can initially look similar.
The company says it used more than 700,000 A100-equivalent GPU hours for automated red teaming focused on universal jailbreaks, alongside human expert and third-party testing. OpenAI plans to keep testing during the preview period and publish an updated system card when the GPT-5.6 family moves toward general availability.
Pricing starts at $5 per 1 million input tokens and $30 per 1 million output tokens for Sol, $2.50 input and $15 output for Terra, and $1 input and $6 output for Luna. GPT-5.6 also adds explicit cache breakpoints, a 30-minute minimum cache life, cache writes at 1.25x the uncached input rate, and cache reads with a 90% cached-input discount. OpenAI also plans to launch GPT-5.6 Sol on Cerebras in July at up to 750 tokens per second for select customers.