Graduate-Level Question Answering on GPQA Diamond
40ScoreEntropy-Cut MH
Evaluation Results
| Method | Links | |
|---|---|---|
| Entropy-Cut MHModel=Qwen3-8B-Base2026.05 | 40 | |
| Uniform-Cut MHModel=Qwen3-8B-Base2026.05 | 38.8 | |
| TMCModel=Qwen3-8B-Base2026.05 | 38.6 | |
| SMCModel=Qwen3-8B-Base2026.05 | 38.1 | |
| Low-TemperatureModel=Qwen3-8B-Base2026.05 | 37.6 | |
| InternVL 3.5 2BNumber of Parameters=2B, Input Modality=Text-Only2026.03 | 34.34 | |
| Entropy-Cut MHModel=Qwen2.5-Math-7B2026.05 | 34.1 | |
| Entropy-Cut MHModel=Phi-4-mini-instruct2026.05 | 33.3 | |
| Low-TemperatureModel=Qwen2.5-Math-7B2026.05 | 33.1 | |
| Xuanwu VL-2BNumber of Parameters=2B, Input Modality=Text-Only2026.03 | 32.83 | |
| Entropy-Cut MHModel=Phi-3.5-mini-instruct2026.05 | 32.8 | |
| TMCModel=Qwen2.5-Math-7B2026.05 | 32.5 | |
| Low-TemperatureModel=Phi-3.5-mini-instruct2026.05 | 32.5 | |
| Uniform-Cut MHModel=Qwen2.5-Math-7B2026.05 | 32.4 | |
| Low-TemperatureModel=Phi-4-mini-instruct2026.05 | 32.1 | |
| TMCModel=Phi-4-mini-instruct2026.05 | 32.1 | |
| SMCModel=Qwen2.5-Math-7B2026.05 | 32 | |
| Uniform-Cut MHModel=Phi-4-mini-instruct2026.05 | 32 | |
| SMCModel=Phi-4-mini-instruct2026.05 | 31.9 | |
| Uniform-Cut MHModel=Phi-3.5-mini-instruct2026.05 | 31.8 | |
| StandardModel=Qwen3-8B-Base2026.05 | 31.7 | |
| TMCModel=Phi-3.5-mini-instruct2026.05 | 31.5 | |
| SMCModel=Phi-3.5-mini-instruct2026.05 | 31 | |
| StandardModel=Phi-4-mini-instruct2026.05 | 30.7 | |
| Entropy-Cut MHModel=Qwen2.5-7B2026.05 | 30.2 | |
| InternVL 3.0 2BNumber of Parameters=2B, Input Modality=Text-Only2026.03 | 29.8 | |
| Low-TemperatureModel=Qwen2.5-7B2026.05 | 29.5 | |
| StandardModel=Qwen2.5-7B2026.05 | 29.4 | |
| Uniform-Cut MHModel=Qwen2.5-7B2026.05 | 29.4 | |
| StandardModel=Qwen2.5-Math-7B2026.05 | 29 | |
| SMCModel=Qwen2.5-7B2026.05 | 28.8 | |
| StandardModel=Phi-3.5-mini-instruct2026.05 | 27.8 | |
| TMCModel=Qwen2.5-7B2026.05 | 27.5 |