Code Generation on LCB (test)
21.69AccuracyPRISM
Evaluation Results
| Method | Links | |
|---|---|---|
| PRISMModel=Qwen2.5-7B, Training Data=DAPO-17k2026.01 | 21.69 | |
| INT.Model=Qwen2.5-7B, Training Data=DAPO-17k2026.01 | 21.67 | |
| BaseModel=Qwen2.5-7B, Training Data=DAPO-17k2026.01 | 21.2 | |
| GRPO (GT)Model=Qwen2.5-7B, Training Data=DAPO-17k2026.01 | 21.17 | |
| PRMModel=Qwen2.5-7B, Training Data=DAPO-17k2026.01 | 21.16 | |
| GRPO (GT)Model=Qwen2.5-3B, Training Data=MATH2026.01 | 14.5 | |
| PRMModel=Qwen2.5-3B, Training Data=MATH2026.01 | 14.5 | |
| PRISMModel=Qwen2.5-3B, Training Data=MATH2026.01 | 14 | |
| BaseModel=Qwen2.5-3B, Training Data=MATH2026.01 | 13.5 | |
| INT.Model=Qwen2.5-3B, Training Data=MATH2026.01 | 12.9 |