Reasoning on O-Bench (test)
56.32AVG@8IAC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| IACModel=Qwen3-14B2026.01 | 56.32 | 62.37 | |
| Reinforce++Model=Qwen3-14B2026.01 | 56 | 61.42 | |
| DrGRPOModel=Qwen3-14B2026.01 | 55.35 | 60.96 | |
| GRPOModel=Qwen3-14B2026.01 | 55.28 | 60.87 | |
| RLOOModel=Qwen3-14B2026.01 | 55.1 | 61.14 | |
| IACModel=Qwen3-8B2026.01 | 50.49 | 58.48 | |
| Reinforce++Model=Qwen3-8B2026.01 | 50.3 | 57.78 | |
| DrGRPOModel=Qwen3-8B2026.01 | 49.87 | 57.26 | |
| GRPOModel=Qwen3-8B2026.01 | 49.79 | 56.88 | |
| RLOOModel=Qwen3-8B2026.01 | 49.74 | 57.68 | |
| IACModel=Qwen3-4B2026.01 | 48.87 | 56.59 | |
| GRPOModel=Qwen3-4B2026.01 | 48.85 | 56.44 | |
| Reinforce++Model=Qwen3-4B2026.01 | 48.57 | 56.38 | |
| RLOOModel=Qwen3-4B2026.01 | 47.91 | 56.31 | |
| DrGRPOModel=Qwen3-4B2026.01 | 47.83 | 55.08 |