Mathematical Reasoning and Coding on DAPO-17k
49.47Peak Accuracy@8seq-agg
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| seq-aggModel=Qwen3-1.7B, Aggregation=seq-agg2026.04 | 49.47 | 59.9 | 44.81 | 55.22 | |
| balanced-aggModel=Qwen3-1.7B, Aggregation=balanced-agg2026.04 | 49.28 | 60.86 | 46.95 | 58.61 | |
| token-aggModel=Qwen3-1.7B, Aggregation=token-agg2026.04 | 48.76 | 60.58 | 43.6 | 56.08 | |
| balanced-aggModel=Qwen2.5-Math-7B, Aggregation=balanced-agg2026.04 | 36.34 | 48.96 | 34.24 | 46.27 | |
| token-aggModel=Qwen2.5-Math-7B, Aggregation=token-agg2026.04 | 35.95 | 48.69 | 33.64 | 44.34 | |
| seq-aggModel=Qwen2.5-Math-7B, Aggregation=seq-agg2026.04 | 35.78 | 48.21 | 34.46 | 45.98 |