Is Raw Qwen3 0.6B direct logits SOTA in Decision models?#448 in Decision modelsRanked #448 in Decision models.
How it scores here
Raw Qwen3 0.6B direct logits in Decision models
Primary benchmarks
- JevBench: 8.2 (CI 7.7–8.7) on Oct 8, 2026 (#133 of 134).
On the category's weighted score it is 99.2 points behind #1 karpo-decision-31B and 0.1 behind the next model up, Julia 1.
Nearby on the board
| # | Model | Category score |
|---|---|---|
| 445 | system-one-gemma | 1.4 |
| 446 | DM-JEPA 1.1 | 1.2 |
| 447 | Julia 1 | 0.9 |
| 448 | Raw Qwen3 0.6B direct logits | 0.8 |
| 449 | Nirnaya 0.4B | 0.6 |
| 450 | Zev | 0.3 |
| 451 | Fastino GLiNER-2.5-Decide (hosted API) | 0.0 |
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 |