GPT-6.1 Sol
OpenAI · Multimodal model
GPT-6.1 Sol is OpenAI's mid-tier model, announced at DevDay 2026 as an upgrade to GPT-6 Sol. It aims to deliver close to Astra-level results for coding and professional work at a much lower price.
Current
ReleasedSep 29, 2026
Context window1.05M tokens (922K max input, 128K max output)
Input → outputtext, image → text
API price$2 / 1M input tokens, $0.10 / 1M cached input tokens, $10 / 1M output tokens
Consumer accessChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu at launch (not initially in the standard Chat mode)
Best for
Complex coding, computer use and professional tasks you want near-Astra quality on at about one-fifth of Astra's token price.
Overview
GPT-6.1 Sol is OpenAI's mid-tier model, announced at DevDay 2026 as an upgrade to GPT-6 Sol. It aims to deliver close to Astra-level results for coding and professional work at a much lower price.
Strengths and limits
- OpenAI says it nearly matches GPT-6 Astra on agentic coding and computer use
- Token prices one-fifth of Astra's; cached input cut to $0.10/M
- Ultrafast speed tier available (from 2026-10-08) in API, Codex and ChatGPT Work
- Tool calling requires the Responses API; no 'none' or 'minimal' reasoning effort
- Some developers report over-refusals on allowed Work/Codex tasks
Published benchmark scores
DeepSWE v1.1 (high effort)75.2%
Source: community.openai.com
OSWorld 2.0 (offline set, max effort)71.4%
Source: community.openai.com
AutomationBench (medium effort)31.7%
Source: community.openai.com
Scores are as published by the source; settings (reasoning effort, tools, attempts) differ between labs, so compare with care. How to read benchmarks →