Mathematical Reasoning on BRUMO 2025 (Accuracy)
93.3AccuracyMajority
Evaluation Results
| Method | Links | |
|---|---|---|
| MajorityModel=DeepSeek-R1-8B2026.05 | 93.3 | |
| DC-MeanModel=DeepSeek-R1-8B2026.05 | 93.3 | |
| DC-TailModel=DeepSeek-R1-8B2026.05 | 93.3 | |
| D-CDG (α = 1)Model=DeepSeek-R1-8B2026.05 | 93.3 | |
| D-CDG (β = 0)Model=DeepSeek-R1-8B2026.05 | 93.3 | |
| CDG (ours)Model=DeepSeek-R1-8B2026.05 | 93.3 | |
| CDG (ours)Model=QwQ-32B2026.05 | 90 | |
| DC-TailModel=QwQ-32B2026.05 | 86.7 | |
| D-CDG (α = 1)Model=QwQ-32B2026.05 | 86.7 | |
| DC-MeanModel=gpt-oss-20B2026.05 | 83.3 | |
| DC-TailModel=gpt-oss-20B2026.05 | 83.3 | |
| D-CDG (α = 1)Model=gpt-oss-20B2026.05 | 83.3 | |
| D-CDG (β = 0)Model=gpt-oss-20B2026.05 | 83.3 | |
| CDG (ours)Model=gpt-oss-20B2026.05 | 83.3 | |
| MajorityModel=gpt-oss-20B2026.05 | 80 | |
| MajorityModel=QwQ-32B2026.05 | 80 | |
| DC-MeanModel=QwQ-32B2026.05 | 80 | |
| D-CDG (β = 0)Model=QwQ-32B2026.05 | 80 | |
| Pass@1Model=DeepSeek-R1-8B2026.05 | 79.7 | |
| Pass@1Model=QwQ-32B2026.05 | 76.8 | |
| Self-Cert.Backbone Model=Qwen3-8B2025.10 | 71 | |
| Loc. uniBackbone Model=Qwen3-8B2025.10 | 70 | |
| Mean Acc.Backbone Model=Qwen3-8B2025.10 | 68 | |
| Glob. uniBackbone Model=Qwen3-8B2025.10 | 68 | |
| Pass@1Model=gpt-oss-20B2026.05 | 64.8 | |
| Glob. non-uniBackbone Model=DS-R1-Distill-Qwen-7B2025.10 | 64 | |
| Low Ent.Backbone Model=Qwen3-8B2025.10 | 64 | |
| High Conf.Backbone Model=Qwen3-8B2025.10 | 63 | |
| Loc. non-uniBackbone Model=Qwen3-8B2025.10 | 63 | |
| Glob. non-uniBackbone Model=Qwen3-8B2025.10 | 61 | |
| Low Ent.Backbone Model=DS-R1-Distill-Qwen-7B2025.10 | 56 | |
| Loc. uniBackbone Model=DS-R1-Distill-Qwen-7B2025.10 | 56 | |
| Mean Acc.Backbone Model=DS-R1-Distill-Qwen-7B2025.10 | 54 | |
| Self-Cert.Backbone Model=DS-R1-Distill-Qwen-7B2025.10 | 52 | |
| High Conf.Backbone Model=DS-R1-Distill-Qwen-7B2025.10 | 52 | |
| DC-TailModel=Gemma-3-27B2026.05 | 46.7 | |
| D-CDG (α = 1)Model=Gemma-3-27B2026.05 | 46.7 | |
| CDG (ours)Model=Gemma-3-27B2026.05 | 46.7 | |
| Glob. uniBackbone Model=DS-R1-Distill-Qwen-7B2025.10 | 44 | |
| Loc. uniBackbone Model=DS-R1-Distill-Llama-8B2025.10 | 44 | |
| D-CDG (β = 0)Model=Gemma-3-27B2026.05 | 43.3 | |
| Mean Acc.Backbone Model=DS-R1-Distill-Llama-8B2025.10 | 40 | |
| Glob. uniBackbone Model=DS-R1-Distill-Llama-8B2025.10 | 40 | |
| MajorityModel=Gemma-3-27B2026.05 | 40 | |
| DC-MeanModel=Gemma-3-27B2026.05 | 40 | |
| Loc. non-uniBackbone Model=DS-R1-Distill-Qwen-7B2025.10 | 39 | |
| Low Ent.Backbone Model=DS-R1-Distill-Llama-8B2025.10 | 37 | |
| High Conf.Backbone Model=DS-R1-Distill-Llama-8B2025.10 | 36 | |
| Loc. non-uniBackbone Model=DS-R1-Distill-Llama-8B2025.10 | 36 | |
| Glob. non-uniBackbone Model=DS-R1-Distill-Llama-8B2025.10 | 36 | |
| Pass@1Model=Gemma-3-27B2026.05 | 35.9 | |
| Self-Cert.Backbone Model=DS-R1-Distill-Llama-8B2025.10 | 34 |