Web Search on HotpotQA (Latency and Cost)
88.5Performance ScoreMasRouter
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MasRouterSetting=Routing2026.01 | 88.5 | 59.5 | 53.1 | |
| EvoRouteSetting=Ours2026.01 | 87.8 | 49.1 | 60.5 | |
| Gemini-2.5-proSetting=Manual2026.01 | 87.4 | 343.43 | 79.18 | |
| GPT-4.1Setting=Manual2026.01 | 87 | 213.22 | 78.77 | |
| GraphRouterSetting=Routing2026.01 | 86.1 | 64.8 | 61.8 | |
| PromptLLMSetting=Routing2026.01 | 85.2 | 47.02 | 62.46 | |
| GPT-4oSetting=Manual2026.01 | 83.84 | 391.98 | 67.29 | |
| Qwen3-14bSetting=Manual2026.01 | 81.4 | 22.96 | 35.52 |