Is Winnow-E4B SOTA in Decision models?#308 in Decision modelsRanked #308 in Decision models.
How it scores here
Winnow-E4B in Decision models
Primary benchmarks
- Decision Index Full: 32.7 (CI 32.2–33.2) on Nov 21, 2026 (#294 of 420).
On the category's weighted score it is 69.0 points behind #1 karpo-decision-31B and 0.0 behind the next model up, Nemotron Diffusion 8B (pst2154, optimized vLLM).
Nearby on the board
| # | Model | Category score |
|---|---|---|
| 305 | LiquidAI d1-Omni-2B | 31.6 |
| 306 | Adjudex-0.8B v6 | 31.3 |
| 307 | Nemotron Diffusion 8B (pst2154, optimized vLLM) | 31.1 |
| 308 | Winnow-E4B | 31.0 |
| 309 | Aurelis-2B | 30.7 |
| 310 | GLiClass v4.1 edge (0.4B) | 30.4 |
| 311 | Qwen3-Reranker-4B | 30.3 |
Standings
Decision models
| # | Model | State |
|---|---|---|
| 1 | karpo-decision-31B | Co-SOTA |
| 1 | Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout) | Co-SOTA |
| 1 | GLiDE 27B | Co-SOTA |
| 1 | kudhab (Gemma 4 31B) | Co-SOTA |
| 2 | Quyet-1.0-Large (Chinh Nguyen, Gemma-4-31B decoder, option-letter logits) | Near-SOTA |
| 3 | REAX 1.3 | Near-SOTA |
| 4 | Logits Sigma-1 Duo | Near-SOTA |
| 5 | KnowLine 31B Gen2 | Near-SOTA |
| 6 | Strands Decider v2 31B | Near-SOTA |
| 7 | Copernicus-31B | Near-SOTA |
| 8 | decisio v0.8.0 on gemma-4-31B-it (frozen, FP8 on load, one prefill per state, self-hosted) | Near-SOTA |
| 9 | Copernicus-57B | Near-SOTA |