Mathematical Reasoning on GSM8K (test) (Accuracy Variants & Relative Change)
77.3Accuracy (Base)R1-Distill-14B
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| R1-Distill-14BBase Model=Qwen2.5-14B2025.10 | 77.3 | 81.7 | 97.2 | 22.2 | |
| R1-Distill-1.5BBase Model=Qwen2.5-Math-1.5B2025.10 | 74 | 73.5 | 90.8 | -3 | |
| ORZ-7BBase Model=Qwen2.5-7B2025.10 | 71.5 | 90.7 | 93.4 | 87.8 | |
| QwQ-32BBase Model=Qwen2.5-32B2025.10 | 69.7 | 77.9 | 97.9 | 29.1 | |
| ORZ-32BBase Model=Qwen2.5-32B2025.10 | 69.5 | 95.5 | 97.2 | 93.6 | |
| R1-Distill-32BBase Model=Qwen2.5-32B2025.10 | 69.4 | 78.4 | 97.6 | 31.8 | |
| ORZ-1.5BBase Model=Qwen2.5-1.5B2025.10 | 51.9 | 68.1 | 74.7 | 71.3 | |
| R1-Distill-8BBase Model=Llama-3.1-8B2025.10 | 45.6 | 44 | 94.3 | -3.4 | |
| ORZ-0.5BBase Model=Qwen2.5-0.5B2025.10 | 25 | 43.2 | 47 | 83 |