General Reasoning on Overall
93.51AccuracyROMCSC
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ROMCSC2026.03 | 93.51 | 947 | 1,159 | 15.57 | 12.37 | |
| EAT2026.03 | 93.47 | 1,535 | 1,843 | 7.57 | 6.26 | |
| L12026.03 | 93.07 | 1,049 | 1,313 | 10.08 | 8.1 | |
| Cut20482026.03 | 92.77 | 1,282 | 1,842 | 7.86 | 5.79 | |
| ROM2026.03 | 92.53 | 1,188 | 1,421 | 11.16 | 9.14 | |
| Vanilla2026.03 | 91.72 | 1,636 | 2,197 | 7.04 | 5.6 | |
| Cut10242026.03 | 91.29 | 912 | 1,743 | 10.15 | 5.9 | |
| Cut5122026.03 | 84.95 | 513 | 1,680 | 16.58 | 5.83 | |
| MCPO-DAPOBackbone=Qwen3-8B2026.05 | 55.87 | — | — | — | — | |
| MCPO-GRPOBackbone=Qwen3-8B2026.05 | 54.22 | — | — | — | — | |
| MCPO-DAPOBackbone=Qwen3-4B2026.05 | 52.6 | — | — | — | — | |
| MT-GRPOBackbone=Qwen3-8B2026.05 | 52.09 | — | — | — | — | |
| MGSBackbone=Qwen3-8B2026.05 | 51.77 | — | — | — | — | |
| MCPO-GRPOBackbone=Qwen3-4B2026.05 | 51.69 | — | — | — | — | |
| DAPOBackbone=Qwen3-8B2026.05 | 51.68 | — | — | — | — | |
| GRPOBackbone=Qwen3-8B2026.05 | 50.51 | — | — | — | — | |
| CLIPOBackbone=Qwen3-8B2026.05 | 50.46 | — | — | — | — | |
| MT-GRPOBackbone=Qwen3-4B2026.05 | 49.37 | — | — | — | — | |
| DAPOBackbone=Qwen3-4B2026.05 | 49.24 | — | — | — | — | |
| MGSBackbone=Qwen3-4B2026.05 | 49.18 | — | — | — | — | |
| GRPOBackbone=Qwen3-4B2026.05 | 47.95 | — | — | — | — | |
| CLIPOBackbone=Qwen3-4B2026.05 | 47.22 | — | — | — | — | |
| Base ModelBackbone=Qwen3-8B2026.05 | 41.34 | — | — | — | — | |
| Base ModelBackbone=Qwen3-4B2026.05 | 37.96 | — | — | — | — |