Math Problem Solving on GSM8K (test)
90.98AccuracyEurus-2-7B-PRIME
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Eurus-2-7B-PRIMEModel Scale=7B2025.05 | 90.98 | 302.72 | |
| SelfBudgeterModel Scale=7B2025.05 | 90.3 | 991.13 | |
| DeepSeek-R1-Distill-QwenModel Scale=7B2025.05 | 87.09 | 1,918.21 | |
| SelfBudgeterModel Scale=1.5B2025.05 | 84.1 | 1,231.79 | |
| L1-MaxSpecified Limit=3600 tokens2025.05 | 79.56 | 571.72 | |
| Qwen-2.5-7B-Simple-RLModel Scale=7B2025.05 | 75.94 | 519.07 | |
| DeepSeek-R1-Distill-QwenModel Scale=1.5B2025.05 | 73.09 | 2,865.08 | |
| E1-Math-1.5BTruncation Budget=4K, 1K2025.05 | 72.1 | 1,299.62 | |
| E1-Math-1.5BTruncation Budget=0.5K, 1K2025.05 | 60.2 | 1,205.21 |