Mathematical Reasoning on MATH L1-L5
92.94Accuracy (L1)AutoGen
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| AutoGenArchitecture=AutoGen, Model=Llama-3.1-70B2026.05 | 92.94 | 82.48 | |
| Multi-Agent DebateArchitecture=Multi-Agent Debate, Model=Llama-3.1-70B2026.05 | 92.94 | 85.9 | |
| Naive-CoTArchitecture=Naive-CoT, Model=Llama-3.1-70B2026.05 | 88.82 | 80.77 | |
| AutoGenArchitecture=AutoGen, Model=Qwen2.5-72B2026.05 | 85.29 | 71.79 | |
| Multi-Agent DebateArchitecture=Multi-Agent Debate, Model=Qwen2.5-72B2026.05 | 79.41 | 66.24 | |
| Naive-CoTArchitecture=Naive-CoT, Model=Qwen2.5-72B2026.05 | 77.65 | 63.68 | |
| ReActArchitecture=ReAct, Model=Qwen2.5-72B2026.05 | 35.88 | 28.21 | |
| ReActArchitecture=ReAct, Model=Llama-3.1-70B2026.05 | 11.76 | 4.27 |