xbench-DeepSearch — StepFun 3.5 Flash card; single-vintage-pinned snapshot, narrower range than the mixed-vintage rollup
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | Step 3.5 Flash | StepFun | 56.3 | 1 | 2026-05-03 |
| 2 | DeepSeek-V3.2 | DeepSeek | 55.7 | 1 | 2026-05-03 |
| 3 | GLM-4.7 | Z.ai | 52.3 | 1 | 2026-05-03 |
| 4 | MiMo V2 Flash | Xiaomi | 44 | 1 | 2026-05-03 |
| 5 | MiniMax M2.1 | MiniMax | 43 | 1 | 2026-05-03 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.