Mathematical Reasoning on AIME 24 (ACC, LEN, delta_LEN)
65.56AccuracyCER
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CERModel=Qwen3-8B2025.10 | 65.56 | 9,171.56 | 18.46 | |
| w thinkingModel=Qwen3-4B2025.10 | 64.44 | 11,343.57 | — | |
| w thinkingModel=Qwen3-8B2025.10 | 63.33 | 11,247.68 | — | |
| CCoTModel=Qwen3-8B2025.10 | 63.33 | 9,286.63 | 17.44 | |
| CERModel=Qwen3-4B2025.10 | 61.11 | 9,215.77 | 18.76 | |
| CCoTModel=Qwen3-4B2025.10 | 56.67 | 9,491.87 | 16.32 | |
| w/o thinkingModel=Qwen3-4B2025.10 | 26.67 | 2,132.2 | 78.34 | |
| CoDModel=Qwen3-4B2025.10 | 23.3 | 3,607.83 | 68.19 | |
| w/o thinkingModel=Qwen3-8B2025.10 | 20 | 2,399.13 | 80.1 | |
| CoDModel=Qwen3-8B2025.10 | 20 | 3,657.57 | 67.48 |