Math Reasoning on IMO-AnswerBench 50
40.5Pass@1 AccuracyTMAS
Evaluation Results
| Method | Links | |
|---|---|---|
| TMASBackbone=Qwen3-30B-Thinking-2507, Iteration=It192026.05 | 40.5 | |
| TMASBackbone=Qwen3-30B-Thinking-2507, Iteration=It172026.05 | 39.31 | |
| RSEBackbone=Qwen3-30B-Thinking-2507, Iteration=It172026.05 | 38.38 | |
| RSEBackbone=Qwen3-30B-Thinking-2507, Iteration=It192026.05 | 38 | |
| TMASBackbone=Qwen3-30B-Thinking-2507, Iteration=It112026.05 | 37.81 | |
| TMASBackbone=Qwen3-30B-Thinking-2507, Iteration=It92026.05 | 36.5 | |
| RSEBackbone=Qwen3-30B-Thinking-2507, Iteration=It112026.05 | 35.31 | |
| RSEBackbone=Qwen3-30B-Thinking-2507, Iteration=It92026.05 | 35.12 | |
| V-RBackbone=Qwen3-30B-Thinking-2507, Iteration=It192026.05 | 31.06 | |
| V-RBackbone=Qwen3-30B-Thinking-2507, Iteration=It172026.05 | 30.88 | |
| TMAS w/ Hybrid-RLBackbone=Qwen3-4B-Thinking-2507, Iteration=It192026.05 | 30.88 | |
| TMAS w/ Hybrid-RLBackbone=Qwen3-4B-Thinking-2507, Iteration=It172026.05 | 30.44 | |
| PaCoReBackbone=Qwen3-30B-Thinking-2507, Iteration=It92026.05 | 30.31 | |
| PaCoReBackbone=Qwen3-30B-Thinking-2507, Iteration=It192026.05 | 30.31 | |
| PaCoReBackbone=Qwen3-30B-Thinking-2507, Iteration=It112026.05 | 30.25 | |
| PaCoReBackbone=Qwen3-30B-Thinking-2507, Iteration=It172026.05 | 29.44 | |
| PaCoReBackbone=Qwen3-30B-Thinking-2507, Iteration=It32026.05 | 29.31 | |
| TMAS w/ Hybrid-RLBackbone=Qwen3-4B-Thinking-2507, Iteration=It92026.05 | 29.25 | |
| TMAS w/ Hybrid-RLBackbone=Qwen3-4B-Thinking-2507, Iteration=It112026.05 | 29.19 | |
| TMASBackbone=Qwen3-30B-Thinking-2507, Iteration=It32026.05 | 28.56 | |
| PaCoReBackbone=Qwen3-30B-Thinking-2507, Iteration=It12026.05 | 26.56 | |
| V-RBackbone=Qwen3-30B-Thinking-2507, Iteration=It112026.05 | 25.75 | |
| RSEBackbone=Qwen3-30B-Thinking-2507, Iteration=It32026.05 | 25.38 | |
| RSEBackbone=Qwen3-30B-Thinking-2507, Iteration=It12026.05 | 25.31 | |
| Self-RefineBackbone=Qwen3-30B-Thinking-2507, Iteration=It192026.05 | 24.19 | |
| V-RBackbone=Qwen3-30B-Thinking-2507, Iteration=It92026.05 | 24.19 | |
| MV@64Backbone=Qwen3-30B-Thinking-2507, Iteration=It12026.05 | 24 | |
| TMAS w/ Hybrid-RLBackbone=Qwen3-4B-Thinking-2507, Iteration=It32026.05 | 22.69 | |
| TMASBackbone=Qwen3-30B-Thinking-2507, Iteration=It12026.05 | 22.06 | |
| Self-RefineBackbone=Qwen3-30B-Thinking-2507, Iteration=It172026.05 | 21.88 | |
| Self-RefineBackbone=Qwen3-30B-Thinking-2507, Iteration=It112026.05 | 18.94 | |
| Self-RefineBackbone=Qwen3-30B-Thinking-2507, Iteration=It92026.05 | 17.5 | |
| TMASBackbone=Qwen3-4B-Thinking-2507, Iteration=It172026.05 | 17.19 | |
| TMASBackbone=Qwen3-4B-Thinking-2507, Iteration=It192026.05 | 17.06 | |
| V-RBackbone=Qwen3-30B-Thinking-2507, Iteration=It32026.05 | 16.31 | |
| RSEBackbone=Qwen3-4B-Thinking-2507, Iteration=It192026.05 | 16.19 | |
| TMASBackbone=Qwen3-4B-Thinking-2507, Iteration=It92026.05 | 15.62 | |
| RSEBackbone=Qwen3-4B-Thinking-2507, Iteration=It172026.05 | 15.44 | |
| TMAS w/ Hybrid-RLBackbone=Qwen3-4B-Thinking-2507, Iteration=It12026.05 | 15.38 | |
| TMASBackbone=Qwen3-4B-Thinking-2507, Iteration=It112026.05 | 14.38 | |
| RSEBackbone=Qwen3-4B-Thinking-2507, Iteration=It32026.05 | 13.31 | |
| RSEBackbone=Qwen3-4B-Thinking-2507, Iteration=It112026.05 | 12.88 | |
| TMASBackbone=Qwen3-4B-Thinking-2507, Iteration=It32026.05 | 12.88 | |
| Self-RefineBackbone=Qwen3-30B-Thinking-2507, Iteration=It32026.05 | 12.75 | |
| V-RBackbone=Qwen3-4B-Thinking-2507, Iteration=It192026.05 | 12.56 | |
| V-RBackbone=Qwen3-4B-Thinking-2507, Iteration=It172026.05 | 12.12 | |
| RSEBackbone=Qwen3-4B-Thinking-2507, Iteration=It12026.05 | 11.38 | |
| PaCoReBackbone=Qwen3-4B-Thinking-2507, Iteration=It172026.05 | 11.12 | |
| RSEBackbone=Qwen3-4B-Thinking-2507, Iteration=It92026.05 | 11.12 | |
| PaCoReBackbone=Qwen3-4B-Thinking-2507, Iteration=It92026.05 | 10.94 | |
| PaCoReBackbone=Qwen3-4B-Thinking-2507, Iteration=It192026.05 | 10.94 | |
| PaCoReBackbone=Qwen3-4B-Thinking-2507, Iteration=It32026.05 | 10.75 | |
| V-RBackbone=Qwen3-30B-Thinking-2507, Iteration=It12026.05 | 10.56 | |
| PaCoReBackbone=Qwen3-4B-Thinking-2507, Iteration=It112026.05 | 10.19 | |
| V-RBackbone=Qwen3-4B-Thinking-2507, Iteration=It112026.05 | 9.69 | |
| Self-RefineBackbone=Qwen3-4B-Thinking-2507, Iteration=It172026.05 | 9.31 | |
| Self-RefineBackbone=Qwen3-4B-Thinking-2507, Iteration=It112026.05 | 9.25 | |
| Self-RefineBackbone=Qwen3-30B-Thinking-2507, Iteration=It12026.05 | 9.06 | |
| V-RBackbone=Qwen3-4B-Thinking-2507, Iteration=It92026.05 | 9.06 | |
| Self-RefineBackbone=Qwen3-4B-Thinking-2507, Iteration=It192026.05 | 8.88 | |
| Self-RefineBackbone=Qwen3-4B-Thinking-2507, Iteration=It92026.05 | 8.31 | |
| V-RBackbone=Qwen3-4B-Thinking-2507, Iteration=It32026.05 | 7.69 | |
| PaCoReBackbone=Qwen3-4B-Thinking-2507, Iteration=It12026.05 | 7.62 | |
| TMASBackbone=Qwen3-4B-Thinking-2507, Iteration=It12026.05 | 6.62 | |
| MV@64Backbone=Qwen3-4B-Thinking-2507, Iteration=It12026.05 | 6 | |
| V-RBackbone=Qwen3-4B-Thinking-2507, Iteration=It12026.05 | 6 | |
| Self-RefineBackbone=Qwen3-4B-Thinking-2507, Iteration=It12026.05 | 5.5 | |
| Self-RefineBackbone=Qwen3-4B-Thinking-2507, Iteration=It32026.05 | 5.44 |