GPT-6.1 Sol
by OpenAI - closed weights. Newest OpenAI model page entry as of 2026-09-30: “near-Astra performance at a lower cost for complex coding, computer use, and professional work”. Hub: gpt-5-and-gpt-6-family.
| Item | Value (OpenAI API docs) |
|---|---|
| API model ID | gpt-6.1-sol |
| Released | 2026-09-29 (API changelog; presented at DevDay per press) |
| Context / output | 1,050,000 window, 922k max input, 128k max output |
| Knowledge cutoff | 2026-04-30 (same as gpt-6-dot-astra; gpt-6-dot-sol is 2026-04-20) |
| Pricing | token-based; cached input is half the price of GPT-6 Sol’s |
| Reasoning effort | low, medium (default), high, xhigh, max (no “none” listed, unlike Sol) |
| Extras | multi-agent functionality in beta (changelog); no Realtime, fine-tuning or embeddings |
What changes vs GPT-6 Sol
Upgrade on coding, computer use and professional work; later knowledge cutoff; cheaper cache reads; same headline price. Benchmark claims are press-reported only and not verified here.
Naming conflict
Press and blogs disagree on whether “6.1” is an official family name (one says unconfirmed); OpenAI’s own docs do use gpt-6.1-sol. Reports say a planned GPT-6.1 Astra was dropped over safety-test results (WSJ/CNBC via search snippets; unverified).
Sources
- https://developers.openai.com/api/docs/models/gpt-6.1-sol (2026-09-30)
- https://developers.openai.com/api/docs/changelog (2026-09-30)
- https://www.developersdigest.tech/blog/gpt-6-1-sol-release-guide-2026 (snippet, 2026-09-30)
- https://www.cnbc.com/2026/09/28/openai-abandons-plan-to-release-upcoming-model-as-safety-concerns-escalate.html (403; snippet only)