Mathematical Reasoning on HMMT (test)
86.9AccuracySelf-Consistency
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Self-ConsistencyModel=Qwen3-next, Family=SC Family2026.06 | 86.9 | — | |
| MARSModel=Qwen3-next, Family=SC Family2026.06 | 86.9 | 31.4 | |
| DeepConfModel=Qwen3-next, Family=DeepConf Family2026.06 | 82.2 | — | |
| MARSModel=Qwen3-next, Family=DeepConf Family2026.06 | 82 | 16.1 | |
| DeepConfModel=DeepSeek-8B, Family=DeepConf Family2026.06 | 78.7 | — | |
| MARSModel=DeepSeek-8B, Family=DeepConf Family2026.06 | 78.6 | 13.8 | |
| Self-ConsistencyModel=DeepSeek-8B, Family=SC Family2026.06 | 70 | — | |
| MARSModel=DeepSeek-8B, Family=SC Family2026.06 | 70 | 29.3 | |
| MARSModel=Qwen3-32B, Family=DeepConf Family2026.06 | 65.5 | 15.7 | |
| DeepConfModel=Qwen3-32B, Family=DeepConf Family2026.06 | 65.1 | — | |
| Self-ConsistencyModel=Qwen3-32B, Family=SC Family2026.06 | 62.8 | — | |
| MARSModel=Qwen3-32B, Family=SC Family2026.06 | 62.8 | 24.7 |