Developer
StepFun
4 indexed models; 1 currently meet the evidence threshold for the overall ranking. Together they have results from 5 evaluations.
Models by StepFun
| Model | Evals | Components | Rank | Release date |
|---|---|---|---|---|
| Step 3.5 Flash | 4 | 5/7 | 203 | 2026-02-01 |
| Step 3 | 2 | 5/7 | — | 2025-07-28 |
| Step 3 Vl 10B | 1 | 1/7 | — | — |
| Step 3.7 Flash | 2 | 2/7 | — | 2026-05-23 |
Evaluations covering StepFun models (5)
AA-Omniscience · BullshitBench v2 · FlagEval Safety and Values · LiveSecBench · TAC