Mathematical Reasoning on AIME (Solved Rate %)
83.33Solved Rate (%)Self-Refine
Evaluation Results
| Method | Links | |
|---|---|---|
| Self-RefineModel=Qwen3 32B think, Context Window=32k2025.05 | 83.33 | |
| ICRLModel=Qwen3 32B think, Context Window=32k2025.05 | 80 | |
| ReflexionModel=Qwen3 32B think, Context Window=32k2025.05 | 70 | |
| BaseModel=Qwen3 32B think, Context Window=32k2025.05 | 66.58 | |
| ICRLModel=Qwen3 32B, Context Window=32k2025.05 | 46.66 | |
| Self-RefineModel=Qwen3 32B, Context Window=32k2025.05 | 43.33 | |
| ReflexionModel=Phi-4, Context Window=16k2025.05 | 40 | |
| ICRLModel=Phi-4, Context Window=16k2025.05 | 40 | |
| ICRLModel=Llama 4 Maverick, Context Window=32k2025.05 | 35 | |
| ReflexionModel=Qwen3 32B, Context Window=32k2025.05 | 33.33 | |
| Self-RefineModel=Phi-4, Context Window=16k2025.05 | 33.33 | |
| ReflexionModel=Llama 4 Maverick, Context Window=32k2025.05 | 23.33 | |
| BaseModel=Qwen3 32B, Context Window=32k2025.05 | 22.54 | |
| Self-RefineModel=Llama 4 Maverick, Context Window=32k2025.05 | 20 | |
| BaseModel=Phi-4, Context Window=16k2025.05 | 20 | |
| BaseModel=Llama 4 Maverick, Context Window=32k2025.05 | 17.58 |