Frontier flagship models
GPT-6.1 Sol
Made by OpenAI
OpenAI's faster, lower-cost GPT-6 model for coding, computer use and professional workflows — released September 2026.
What it is
GPT-6.1 Sol is OpenAI's September 29, 2026 update to the GPT-6 family. It is built for complex coding, computer use, document work and multistep agent workflows, with capabilities OpenAI describes as comparable to GPT-6 Astra at one-fifth of Astra's standard input and output token prices.
The model accepts text and images, produces text, and has a 1,050,000-token context window with up to 128,000 output tokens. The API supports low, medium, high, xhigh and max reasoning effort. Audio and video are not supported.
At launch, it is available through the OpenAI API and in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu users. OpenAI says it is not yet available in the regular ChatGPT chat surface, and API use is billed separately from a ChatGPT subscription.
OpenAI classifies GPT-6.1 Sol as Critical for cybersecurity and High for biological and chemical capability under its Preparedness Framework. The company says it uses the same safeguard stack as GPT-6 Astra.
OpenAI classifies the model as Critical for cybersecurity and High for biological and chemical capability under its Preparedness Framework, and applies the same safeguard stack used for GPT-6 Astra.
Published price
- API — input$2 per 1M tokens
- API — cached input$0.10 per 1M tokens
- API — output$10 per 1M tokens
- Batch50% of standard rates
- ChatGPT Work and CodexPlus, Pro, Business, Enterprise and Edu; not yet available in regular ChatGPT chat at launchIncluded on eligible paid plans
API use is billed separately from a ChatGPT subscription. Tool calling requires the Responses API; Chat Completions supports text-only use.
Source: OpenAI's model documentation · Last checked 29 September 2026
Benchmarks
The published figures for GPT-6.1 Sol, copied exactly as their sources give them and checked 29 September 2026. Each one names the test, who ran it and when, and says when the number comes from the maker rather than an independent tester.
- DeepSWE v1.175.2%
share of software-engineering tasks completed successfully · tested: high reasoning effort
OpenAI · 29 September 2026 · maker's own figures
OpenAI reports 74.1% for GPT-6 Astra in the same launch comparison.
- OSWorld 2.0 (offline set)71.4%
share of computer-use tasks completed successfully · tested: max reasoning effort
OpenAI · 29 September 2026 · maker's own figures
OpenAI reports 73.5% for GPT-6 Astra in the same launch comparison.
These launch figures come from OpenAI's own evaluations. Independent leaderboards may test different settings and production systems.
Figures copied from the sources above, checked 29 September 2026.
Where to read today's numbers
Scores move, so these are the boards that keep them current.
- LMArena leaderboardIndependent
Ranks models by blind head-to-head votes from the public. Good for a feel for general quality, weak on specialist work.
- Artificial AnalysisIndependent
Independent testing of quality, speed and price per million tokens across hosted models.
- Vals AI benchmarksIndependent
Independent evaluations on real professional tasks — law, finance, tax, medicine — rather than vendor demos.
- OpenAI's GPT-6.1 Sol system card addendumMaker's own figures
OpenAI's own safety evaluations and launch comparisons. These are maker-reported results, not independent tests.
How it performs, in my experience
OpenAI's September 29 update to the Sol line, aimed at complex coding, computer use and professional workflows. OpenAI says it approaches GPT-6 Astra's capabilities while charging one-fifth of Astra's standard input and output token rates; the launch benchmark figures are the maker's own.
Prompting it from this site
Not wired up here yet
This site's OpenAI choices still use GPT-5.6 Sol and GPT-6 Astra. Use OpenAI's API, ChatGPT Work or Codex for GPT-6.1 Sol.
Use it for
- Complex coding and debugging
- Computer-use and multistep agent workflows
- Long document and professional production tasks