Mathematical Reasoning on Olympiad (Pass@16)
90.95Pass@16Base
Evaluation results
| Method | Links | |
|---|---|---|
| BaseBase Model=Qwen3-4B-Thinking2026.01 | 90.95 | |
| RF-BBase Model=Qwen3-4B-Thinking2026.01 | 89.32 | |
| DAPOBase Model=Qwen3-4B-Thinking2026.01 | 89.02 | |
| SPAEBase Model=Qwen3-4B-Thinking2026.01 | 89.02 | |
| SPAEBase Model=DeepSeek-R1-Distill-Qwen-7B2026.01 | 88.13 | |
| BaseBase Model=DeepSeek-R1-Distill-Qwen-7B2026.01 | 87.98 | |
| RF-BBase Model=DeepSeek-R1-Distill-Qwen-7B2026.01 | 87.98 | |
| DASTBase Model=DeepSeek-R1-Distill-Qwen-7B2026.01 | 87.83 | |
| EntropyBase Model=DeepSeek-R1-Distill-Qwen-7B2026.01 | 87.83 | |
| DAPOBase Model=DeepSeek-R1-Distill-Qwen-7B2026.01 | 87.69 | |
| LC-R1Base Model=DeepSeek-R1-Distill-Qwen-7B2026.01 | 87.39 | |
| DAPOBase Model=DeepSeek-R1-Distill-Llama-8B2026.01 | 87.39 | |
| KTAEBase Model=DeepSeek-R1-Distill-Qwen-7B2026.01 | 86.8 | |
| RF-BBase Model=DeepSeek-R1-Distill-Llama-8B2026.01 | 86.8 | |
| BaseBase Model=DeepSeek-R1-Distill-Llama-8B2026.01 | 86.5 | |
| SPAEBase Model=DeepSeek-R1-Distill-Llama-8B2026.01 | 86.5 |