Mathematical Reasoning on AIME 2024 (Accuracy and Length Metrics)
53AccuracyCyclicReflex
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CyclicReflexBase Model=DeepSeek-R1-Distill-Llama-8B2025.06 | 53 | 3,454.97 | |
| CyclicReflexBase Model=DeepSeek-R1-Distill-Qwen-7B2025.06 | 50 | 2,868.3 | |
| TIPBase Model=DeepSeek-R1-Distill-Llama-8B2025.06 | 47 | 3,572.4 | |
| OriginalBase Model=DeepSeek-R1-Distill-Qwen-7B2025.06 | 43 | 2,878.39 | |
| TIPBase Model=DeepSeek-R1-Distill-Qwen-7B2025.06 | 43 | 2,806.53 | |
| S1Base Model=DeepSeek-R1-Distill-Llama-8B2025.06 | 43 | 3,742.27 | |
| OriginalBase Model=DeepSeek-R1-Distill-Llama-8B2025.06 | 42 | 3,593.73 | |
| S1Base Model=DeepSeek-R1-Distill-Qwen-7B2025.06 | 33 | 3,541.1 | |
| CyclicReflexBase Model=DeepSeek-R1-Distill-Qwen-1.5B2025.06 | 30 | 3,547.1 | |
| OriginalBase Model=DeepSeek-R1-Distill-Qwen-1.5B2025.06 | 23 | 3,584.36 | |
| TIPBase Model=DeepSeek-R1-Distill-Qwen-1.5B2025.06 | 23 | 3,329.17 | |
| S1Base Model=DeepSeek-R1-Distill-Qwen-1.5B2025.06 | 17 | 4,112.07 |