CONFIRMEDDate confidence: CONFIRMEDStated as firmHow firmly the source words the date: confirmed, scheduled, estimated, delayed or unverified.Show all CONFIRMED pins91% CONFIDENCEOverall confidence: 91%How well the pin's source and references back up its dates.Weighted average of how firmly 3 references, the source included, support the pin's start and end times; a reference counts half as much for every 180 days older than the newestShow all pins at 75% confidence or better
OpenAI's[1][2][3] own model release notes date this entry July 9, 2026- Start
- Jul 9, 202695% CONFIDENCEhelp.openai.com
GPT-5.6 Sol Rolls Out in ChatGPT
OpenAI begins rolling out GPT-5.6 Sol, its flagship reasoning model for complex work across coding, research, science, cybersecurity, computer use and design.
- GPT-5.6 Sol began rolling out in ChatGPT on July 9, 2026 as OpenAI's flagship reasoning model for complex work across coding, research, science, cybersecurity, computer use and design[1]
- Rollout is to eligible paid plans; Free, Go and logged-out users are not included[1]
- The GPT-5.6 series (Sol, Terra, Luna) became generally available on July 9, 2026 at 10am PT, after a preview from June 26 for select trusted partners via API and Codex[3]
Notable features
- Sol is the newest flagship for frontier reasoning and long-horizon agentic work; Terra is a balanced everyday model with GPT-5.5-competitive performance at 2x lower cost; Luna is the fastest and most affordable[3]
- Specifications: a 1,050,000-token context window and 128,000 max output tokens, text and image input with text output, a February 16, 2026 knowledge cutoff, and reasoning effort settings of none, low, medium (the default), high, xhigh and max; the gpt-5.6 API alias routes to Sol[2]
- API pricing is $4 per million input tokens ($0.40 cached) and $20 per million output tokens, which OpenAI calls a 20% cut in input and 33% cut in output pricing, promotional through at least November 21, 2026; prompts over 272K input tokens are billed at 2x input and 1.5x output[2]
- Sol sets a new state of the art on Terminal-Bench 2.1 and adds a "max" reasoning effort and an "ultra mode" using subagents[3]
- On SecureBio evaluations GPT-5.6 reached 53.5% on the Virology Capabilities Test, 60.0% on Molecular Biology, 68.4% on Human Pathogen Capabilities and 68.3% on World-Class Bio, about 9 points above GPT-5.5; on ExploitBench Sol was competitive while using roughly a third of the output tokens of another leading frontier system[3]
Buy onChatGPT