OpenAI publicly launched its most powerful AI model family to date on July 9, releasing the GPT-5.6 series — Sol, Terra, and Luna — alongside a new agentic product called ChatGPT Work, in what analysts are calling the company's biggest product day of the year.
The launch, delayed by roughly two weeks at the request of the U.S. government over national security concerns, marks a major escalation in the AI arms race between OpenAI and rival Anthropic.
Three Models for Three Needs
GPT-5.6 Sol is the flagship: a frontier model achieving state-of-the-art results across coding, cybersecurity, scientific research, and knowledge work. On the Artificial Analysis Coding Agent Index, Sol with max reasoning scored 80 — 2.8 points above Claude Fable 5 — while using less than half the output tokens at roughly a third less cost. On the Agents' Last Exam benchmark across 55 fields, Sol reached 53.6, eclipsing Fable 5 by 13.1 points.
Terra is a balanced, lower-cost model with performance competitive with GPT-5.5. Luna is the fastest, most cost-efficient option at just $1 input and $6 output per million tokens.
ChatGPT Work: The Super App
OpenAI also unveiled ChatGPT Work, an AI agent combining ChatGPT with Codex to create documents, presentations, and websites from natural-language prompts. Designed for non-coders, it directly competes with Anthropic's Claude Cowork.
"You can apply the model's ability to code to solve problems across every industry," said Ty Geri, product manager for ChatGPT Work.
Safety and Pricing
GPT-5.6 ships with OpenAI's most robust safety stack, backed by ~700,000 A100e GPU hours of automated red-teaming. Pricing ranges from $5/$30 per million tokens (Sol) down to $1/$6 (Luna). Rollout began July 9 and continues over 24 hours.
