Mathematical Reasoning on AIME 2025 (Acc, Len)
37AccuracyCyclicReflex
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CyclicReflexBase Model=DeepSeek-R1-Distill-Qwen-7B2025.06 | 37 | 3,190.33 | |
| CyclicReflexBase Model=DeepSeek-R1-Distill-Llama-8B2025.06 | 37 | 3,856.8 | |
| S1Base Model=DeepSeek-R1-Distill-Qwen-7B2025.06 | 33 | 3,455.33 | |
| OriginalBase Model=DeepSeek-R1-Distill-Qwen-7B2025.06 | 31 | 3,192.59 | |
| TIPBase Model=DeepSeek-R1-Distill-Qwen-7B2025.06 | 30 | 3,107.3 | |
| OriginalBase Model=DeepSeek-R1-Distill-Llama-8B2025.06 | 30 | 3,922.41 | |
| TIPBase Model=DeepSeek-R1-Distill-Llama-8B2025.06 | 27 | 3,866 | |
| S1Base Model=DeepSeek-R1-Distill-Llama-8B2025.06 | 27 | 4,351.87 | |
| CyclicReflexBase Model=DeepSeek-R1-Distill-Qwen-1.5B2025.06 | 23 | 3,467.97 | |
| TIPBase Model=DeepSeek-R1-Distill-Qwen-1.5B2025.06 | 20 | 3,825.17 | |
| S1Base Model=DeepSeek-R1-Distill-Qwen-1.5B2025.06 | 20 | 3,867.71 | |
| OriginalBase Model=DeepSeek-R1-Distill-Qwen-1.5B2025.06 | 19 | 3,442.07 |