| Benchmark | Model | Rank | Results |
|---|---|---|---|
| code-generation-on-humaneval | ERC-DAPO (DeepSeek-R1-Distill-Qwen-7B) | – | Pass@1: 91.3 |
| mathematical-reasoning-on-aime-2025 | ERC-DAPO (DeepSeek-R1-Distill-Qwen-7B) | – | Accuracy: 48.4 |
| mathematical-reasoning-on-aime24 | ERC-GPPO (DeepSeek-R1-Distill-Qwen-7B) | – | Acc: 63.5 |
| mathematical-reasoning-on-math-500 | ERC-DAPO (DeepSeek-R1-Distill-Qwen-7B) | – | Accuracy: 95.1 |