Next-token reasoning on OMNI-MATH Easy (val)
76.89AccuracyLoopRPT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LoopRPTModel=Ouro-2.6B, Reasoning Mode=Peak2026.03 | 76.89 | 4 | |
| LoopRPTModel=Ouro-2.6B, Reasoning Mode=Adap.2026.03 | 76.07 | 2.05 | |
| LoopRPTModel=Ouro-1.4B, Reasoning Mode=Peak2026.03 | 75.38 | 4 | |
| LoopRPTModel=Ouro-1.4B, Reasoning Mode=Adap.2026.03 | 75 | 2.5 | |
| PeakModel=Ouro-1.4B2026.03 | 74.62 | 4 | |
| Adap.Model=Ouro-2.6B2026.03 | 74.51 | 3.24 | |
| Adap.Model=Ouro-1.4B2026.03 | 74.4 | 3.34 | |
| PeakModel=Ouro-2.6B2026.03 | 74.33 | 4 | |
| VanillaModel=Qwen3-1.7B2026.03 | 47.49 | — | |
| +CoTModel=Qwen3-1.7B2026.03 | 10.7 | — |