Is Raw Qwen3 8B direct logits SOTA in Decision models?#414 in Decision modelsRanked #414 in Decision models.
How it scores here
Raw Qwen3 8B direct logits in Decision models
Primary benchmarks
- JevBench: 30.1 (CI 29.6–30.6) on Oct 8, 2026 (#123 of 134).
On the category's weighted score it is 91.7 points behind #1 karpo-decision-31B and 0.1 behind the next model up, Prumo.
Nearby on the board
| # | Model | Category score |
|---|---|---|
| 411 | Crivo-32M v2 | 9.0 |
| 412 | Crivo-Prumo 17M | 8.7 |
| 413 | Prumo | 8.4 |
| 414 | Raw Qwen3 8B direct logits | 8.3 |
| 415 | Kodiak v0.2 1B (accuracy) | 8.1 |
| 416 | Laya (Convai Innovations, ModernBERT-large 421M) | 8.1 |
| 417 | system-one (Qwen3-8B, Sean Goedecke) | 7.6 |
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 |