Mathematical Reasoning on AMC 2023 (Acc, Len, CR)
97.5AccuracyASAG
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ASAGModel=Qwen3-32B2026.06 | 97.5 | 4,830 | 62.6 | |
| ASAGModel=Qwen3-14B2026.06 | 95 | 4,389 | 60.8 | |
| VanillaModel=Qwen3-32B2026.06 | 95 | 7,711 | 100 | |
| ASAGModel=Qwen3-8B2026.06 | 92.5 | 4,285 | 53.3 | |
| VanillaModel=Qwen3-14B2026.06 | 92.5 | 7,214 | 100 | |
| ASAGModel=Qwen3-4B2026.06 | 90 | 5,230 | 69.6 | |
| VanillaModel=Qwen3-4B2026.06 | 87.5 | 7,512 | 100 | |
| VanillaModel=Qwen3-8B2026.06 | 87.5 | 8,045 | 100 | |
| ASAGModel=DeepSeek-R1-Distill-Qwen-7B2026.06 | 82.5 | 4,312 | 63 | |
| ASAGModel=DeepSeek-R1-Distill-Llama-8B2026.06 | 80 | 4,087 | 63.1 | |
| VanillaModel=DeepSeek-R1-Distill-Qwen-7B2026.06 | 77.5 | 6,841 | 100 | |
| VanillaModel=DeepSeek-R1-Distill-Llama-8B2026.06 | 72.5 | 6,482 | 100 |