StepFun
Step 3.7 Flash
StepFun’s flagship multimodal reasoning model. Powered by a 198B-parameter / 11B-activation sparse MoE architecture, with native support for image and video understanding.
StepFun researchModel Snapshot
TypeLanguage
ReleaseMay 28, 2026
Context256,000
Tagsreasoning, vision, tool-use
Overall
Rank #67 across 523
42.6
Coding
Rank #72 across 433
37.1
Blended Price
Rank #126 across 355
$0.438/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 Speed385.5 tok/s
First Token0.74s
Max Output256,000
Pricing
Blended$0.438/M
Input$0.200/M
All Benchmarks
| Metric | Domain | Value | Rank |
|---|---|---|---|
| Artificial Analysis Intelligence Index | overall | 42.6 | #67 |
| Artificial Analysis Coding Index | coding | 37.1 | #72 |
| GPQA | reasoning | 80.9% | #110 |
| Humanity's Last Exam | reasoning | 19.9% | #80 |
| SciCode |