When was GPT-5.6 released?

OpenAI previewed GPT-5.6 on June 26, 2026 as three models — Sol, Terra, and Luna — through a limited Codex and API preview. On July 9, 2026, @OpenAI posted that Sol, Terra, and Luna are starting to roll out in ChatGPT, Codex, and the API with GA benchmarks including ALE 53.6 and AA Coding Agent Index 80.0.

What are GPT-5.6 Sol, Terra, and Luna?

Sol is OpenAI's GPT-5.6 flagship; Terra delivers GPT-5.5-competitive performance at roughly 2× lower cost; Luna is the fastest and most affordable tier for high-volume work. The 5.6 number marks the generation; Sol/Terra/Luna are durable capability tiers.

How does GPT-5.6 compare to Claude Fable 5?

OpenAI's July 9 GA claims: Sol leads Agents' Last Exam at 53.6 (+13.1 vs Fable adaptive), Artificial Analysis Coding Agent Index at 80.0 (+2.8 vs Fable), and Terminal-Bench 2.1 at 91.9% (Sol Ultra). Fable 5 leads SWE-Bench Pro at 80.3%. Both are publicly accessible as of July 2026.

What improvements does GPT-5.6 Sol make over GPT-5.5?

OpenAI calls Sol a step-function improvement over GPT-5.5 — new max reasoning effort, Ultra mode with subagents, state-of-the-art Terminal-Bench 2.1 scores, stronger biology (GeneBench v1) and cyber evals, and a layered safeguard stack. Terra and Luna trade capability for cost within the same generation.

How much does GPT-5.6 cost?

Official API pricing per 1M tokens: Sol $5/$30 (same list as GPT-5.5), Terra $2.50/$15, Luna $1/$6. GPT-5.6 adds explicit cache breakpoints and 30-minute minimum cache life.

What are Polymarket odds on GPT-5.6 release date?

On July 7, 2026, Polymarket posted ~80% odds on July 10 — a prediction market, not OpenAI. On July 8, 2026, @OpenAI officially said Sol, Terra, and Luna launch publicly this Thursday (July 9) with global preview expansion — superseding Polymarket as the primary schedule signal. Treat markets as crowd sentiment only.

← Back to blog

explainx / blog

GPT-5.6 Guide: Sol, Terra, Luna Models, Pricing, and Benchmarks

OpenAI July 9: GPT-5.6 Sol, Terra, Luna rolling out in ChatGPT, Codex, and API. ALE 53.6, AA Coding Index 80.0, Ultra mode, Terminal-Bench 91.9%, pricing.

Jun 15, 2026·12 min read·Yash Thakker

AIOpenAIGPTLLMFrontier AI