Mathematical Reasoning on AIME 24 (Accuracy, Length)
53.33AccuracyEffRea
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| EffReaBackbone=DeepSeek-R1-7B2026.02 | 53.33 | 10,706.94 | |
| StandardpBackbone=DeepSeek-R1-7B2026.02 | 51.11 | 14,061.22 | |
| LASERDEBackbone=DeepSeek-R1-7B2026.02 | 51.11 | 5,985.25 | |
| C3oTBackbone=DeepSeek-R1-7B2026.02 | 51.11 | 12,884.91 | |
| S3-CoTBackbone=DeepSeek-R1-7B2026.02 | 51.11 | 12,217.53 | |
| S3-CoTscBackbone=DeepSeek-R1-7B2026.02 | 50 | 12,167.95 | |
| EfficientpBackbone=DeepSeek-R1-7B2026.02 | 47.78 | 11,224.51 | |
| AutoTHINKBackbone=DeepSeek-R1-7B2026.02 | 47.78 | 8,303.59 | |
| LC-R1Backbone=DeepSeek-R1-7B2026.02 | 42.22 | 6,371.22 | |
| ShorterBetterBackbone=DeepSeek-R1-7B2026.02 | 37.78 | 4,701.88 | |
| CoT-ValveBackbone=DeepSeek-R1-7B2026.02 | 20 | 4,177.39 | |
| S3-CoTBackbone=Qwen2.5-7B2026.02 | 12.22 | 800.62 | |
| S3-CoTscBackbone=Qwen2.5-7B2026.02 | 11.11 | 831.47 | |
| StandardpBackbone=Qwen2.5-7B2026.02 | 7.78 | 996.75 | |
| EfficientpBackbone=Qwen2.5-7B2026.02 | 7.78 | 741.21 | |
| CoT-ValveBackbone=Qwen2.5-7B2026.02 | 6.68 | 1,068.58 | |
| C3oTBackbone=Qwen2.5-7B2026.02 | 5.56 | 866.8 | |
| TokenSkipBackbone=Qwen2.5-7B2026.02 | 2.33 | 842.71 |