Mathematical Reasoning on AIME 25 (Acc., Tokens)
86.4Accuracy (AIME 25)D-L
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| D-LBackbone=DeepSeek-R1-0528-Qwen3-8B, Voting size budget=5122026.05 | 86.4 | 12.4 | |
| DDCBackbone=DeepSeek-R1-0528-Qwen3-8B, Voting size budget=5122026.05 | 83.3 | 2.4 | |
| SCBackbone=DeepSeek-R1-0528-Qwen3-8B, Voting size budget=5122026.05 | 82.3 | 40.1 | |
| ACBackbone=DeepSeek-R1-0528-Qwen3-8B, Voting size budget=5122026.05 | 82.3 | 13.3 | |
| DDCBackbone=Qwen3-4B, Voting size budget=5122026.05 | 82.1 | 1.5 | |
| D-HBackbone=DeepSeek-R1-0528-Qwen3-8B, Voting size budget=5122026.05 | 81.4 | 23.7 | |
| D-LBackbone=Qwen3-32B, Voting size budget=5122026.05 | 80.2 | 11.4 | |
| D-HBackbone=Qwen3-32B, Voting size budget=5122026.05 | 80.2 | 16.1 | |
| SCBackbone=Qwen3-32B, Voting size budget=5122026.05 | 80.1 | 24.3 | |
| ACBackbone=Qwen3-32B, Voting size budget=5122026.05 | 80 | 9.1 | |
| DDCBackbone=Qwen3-32B, Voting size budget=5122026.05 | 79.7 | 1.8 | |
| D-HBackbone=Qwen3-8B, Voting size budget=5122026.05 | 78.5 | 19.9 | |
| SCBackbone=Qwen3-8B, Voting size budget=5122026.05 | 78.1 | 27.7 | |
| D-LBackbone=Qwen3-8B, Voting size budget=5122026.05 | 76.7 | 13.1 | |
| D-LBackbone=Qwen3-4B, Voting size budget=5122026.05 | 73.3 | 18.3 | |
| DDCBackbone=Qwen3-8B, Voting size budget=5122026.05 | 73.3 | 2.2 | |
| ACBackbone=Qwen3-8B, Voting size budget=5122026.05 | 70 | 9.2 | |
| D-HBackbone=Qwen3-4B, Voting size budget=5122026.05 | 66.7 | 25.8 | |
| SCBackbone=Qwen3-4B, Voting size budget=5122026.05 | 66.5 | 41 | |
| ACBackbone=Qwen3-4B, Voting size budget=5122026.05 | 66.4 | 15.6 | |
| D-LBackbone=Qwen3-1.7B, Voting size budget=5122026.05 | 53.3 | 9.3 | |
| D-HBackbone=Qwen3-1.7B, Voting size budget=5122026.05 | 53 | 23.4 | |
| DDCBackbone=Qwen3-1.7B, Voting size budget=5122026.05 | 50 | 3.7 | |
| SCBackbone=Qwen3-1.7B, Voting size budget=5122026.05 | 49.5 | 47.5 | |
| ACBackbone=Qwen3-1.7B, Voting size budget=5122026.05 | 49.4 | 17.8 |