Math & Physics Reasoning Accuracy on AIME 2024
19.48AccuracyAZR
Evaluation Results
| Method | Links | |
|---|---|---|
| AZRModel Backbone=Qwen3-8B-Base2026.06 | 19.48 | |
| INFUSERModel Backbone=Qwen3-8B-Base2026.06 | 18.58 | |
| SPICEModel Backbone=Qwen3-8B-Base2026.06 | 18.4 | |
| R-ZeroModel Backbone=Qwen3-8B-Base2026.06 | 13.96 | |
| AZRModel Backbone=Qwen3-4B-Base2026.06 | 13.54 | |
| BaseModel Backbone=Qwen3-8B-Base2026.06 | 12.92 | |
| SPICEModel Backbone=Qwen3-4B-Base2026.06 | 12.2 | |
| INFUSERModel Backbone=Qwen3-4B-Base2026.06 | 11.35 | |
| BaseModel Backbone=Qwen3-4B-Base2026.06 | 10.42 | |
| R-ZeroModel Backbone=Qwen3-4B-Base2026.06 | 9.38 |