Mathematical Reasoning on AIME Round 2 2024
0.94Average AccuracyQwen2.5-72B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen2.5-72BBenchmark Version=Original2025.10 | 0.94 | 19,265 | |
| Qwen2.5-72BBenchmark Version=Evolved2025.10 | 0.9033 | 21,808.9 | |
| DeepSeek-R1-Distill-Qwen-32BBenchmark Version=Original2025.10 | 0.73 | 9,912.8 | |
| DeepSeek-R1-Distill-Qwen-32BBenchmark Version=Evolved2025.10 | 0.6233 | 10,556.7 | |
| DeepSeek-R1-Distill-Qwen-7BBenchmark Version=Original2025.10 | 0.5667 | 16,890.5 | |
| DeepSeek-R1-Distill-Qwen-7BBenchmark Version=Evolved2025.10 | 0.4933 | 14,853.4 | |
| Qwen2.5-72BBenchmark Version=Delta2025.10 | -0.0367 | 2,543.9 | |
| DeepSeek-R1-Distill-Qwen-7BBenchmark Version=Delta2025.10 | -0.0734 | -2,037.1 | |
| DeepSeek-R1-Distill-Qwen-32BBenchmark Version=Delta2025.10 | -0.1067 | 643.2 |