OpenAI
GPT-6 Sol (medium)
GPT-6 Sol is OpenAI's reasoning model for complex coding and agentic workflows. It accepts text and images, generates text, and supports a 1,050,000-token context window with up to 128,000 output tokens. Configurable reasoning effort, structured outputs, prompt caching, and Responses API tools make it suited to software development and multi-step automation.
OpenAI model releasesModel Snapshot
TypeLanguage
ReleaseSep 22, 2026
Context1,050,000
Tagsexplicit-caching, fast, file-input, implicit-caching
Overall
Rank #55 across 687
39.8
Coding
No rank
n/a
Blended Price
Rank #343 across 432
$4.00/M
Strength Profile
Percentile score by analysis domain.
* Cost is inverted: lower input, output, and blended prices rank higher.
Benchmark Percentiles
Higher bars mean stronger relative placement.
Model Details
Runtime
Output Speed134.4 tok/s
First Token6.70s
Max Output128,000
Pricing
Blended$4.00/M
Input$2.00/M
All Benchmarks
| Metric | Domain | Value | Rank |
|---|---|---|---|
| Artificial Analysis Intelligence Index | overall | 39.8 | #55 |
| Humanity's Last Exam | reasoning | 41.0% | #64 |
| SciCode | coding, reasoning | 53.8% | #59 |
| Output Speed | speed | 134.4 tok/s | #64 |
| Time to First Token |