Math & physics reasoning on MATH 500
82.77AccuracyINFUSER
Evaluation Results
| Method | Links | |
|---|---|---|
| INFUSERModel Backbone=Qwen3-8B-Base2026.06 | 82.77 | |
| R-FewModel Backbone=Qwen3-8B-Base2026.06 | 82.6 | |
| AZRModel Backbone=Qwen3-8B-Base2026.06 | 80.95 | |
| R-ZeroModel Backbone=Qwen3-8B-Base2026.06 | 80.55 | |
| SPICEModel Backbone=Qwen3-8B-Base2026.06 | 79.4 | |
| R-FewModel Backbone=Qwen3-4B-Base2026.06 | 78 | |
| SPICEModel Backbone=Qwen3-4B-Base2026.06 | 78 | |
| R-ZeroModel Backbone=Qwen3-4B-Base2026.06 | 76.85 | |
| INFUSERModel Backbone=Qwen3-4B-Base2026.06 | 76.65 | |
| BaseModel Backbone=Qwen3-8B-Base2026.06 | 76.05 | |
| AZRModel Backbone=Qwen3-4B-Base2026.06 | 73.9 | |
| BaseModel Backbone=Qwen3-4B-Base2026.06 | 61.2 |