OpenAI Previews GPT-5.6 Sol: A New Flagship Model Family

OpenAI unveiled the GPT-5.6 model family on June 26, introducing three tiers: Sol (flagship), Terra (mid-tier), and Luna (fast/affordable). Sol is positioned as a next-generation reasoning model with particular strength in coding, cybersecurity, and biology, as well as sustained focus during long-horizon agentic tasks.

Sol is priced at $5 per million input tokens and $30 per million output tokens — roughly half the cost of Anthropic’s Claude Fable 5 — with Terra at half Sol’s price and Luna lower still. The flagship also introduces “max” and “ultra” inference modes for deeper reasoning and sub-agent orchestration, respectively.

The preview comes with a notably cautious safety posture. OpenAI reports dedicating approximately 700,000 A100e GPU hours to automated red-teaming and says Sol features the company’s “most robust safety stack to date,” with strengthened protections against jailbreaking, sensitive cyber requests, and repeated misuse. The administration is closely monitoring the preview period, approving customers on a case-by-case basis.

Why it matters: GPT-5.6 Sol marks an inflection point where capability gains are narrowing between frontier labs, pushing competition toward pricing and safety posture rather than raw benchmark supremacy.