Code Generation on Code-Contest
20.3Pass@1JPlag-RLVR
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| JPlag-RLVRModel=QWEN3-8B2026.05 | 20.3 | 35.8 | 47.7 | 46.6 | 41.1 | 2.39 | |
| JPlag-RLVRModel=QWEN3-4B2026.05 | 17.8 | 35.3 | 46.2 | 52.3 | 46 | 2.92 | |
| Base-RLVRModel=QWEN3-8B2026.05 | 17.4 | 33.5 | 45.6 | 37 | 34.8 | 2.37 | |
| Base-RLVRModel=QWEN3-4B2026.05 | 16.8 | 31.5 | 39.6 | 39.6 | 37.5 | 2.32 | |
| PKPOModel=QWEN3-8B2026.05 | 15 | 32 | 45.6 | 41 | 36.6 | 2.43 | |
| Base-RLVRModel=OLMO3-7B2026.05 | 14.5 | 29 | 41.2 | 46.5 | 43.8 | 3.2 | |
| JPlag-RLVRModel=OLMO3-7B2026.05 | 14.3 | 30.1 | 41.2 | 49.5 | 47.4 | 3.33 | |
| Pass@k-RLVRModel=QWEN3-8B2026.05 | 13.5 | 33.9 | 47.6 | 49.5 | 45.5 | 3.39 | |
| PKPOModel=QWEN3-4B2026.05 | 13.4 | 31.7 | 46 | 41 | 35 | 2.4 | |
| Pass@k-RLVRModel=QWEN3-4B2026.05 | 12.1 | 29.3 | 41.4 | 57.2 | 48.5 | 3.29 | |
| PKPOModel=OLMO3-7B2026.05 | 11.9 | 30.5 | 45 | 51.6 | 49.6 | 3.58 | |
| Pass@k-RLVRModel=OLMO3-7B2026.05 | 7.4 | 20.1 | 35 | 53.6 | 55.4 | 4.99 |