Medical Reasoning on DDXPlus
90.2Performance ScoreACE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ACEBase LLM=DeepSeek-V3.1-671B, Adaptation Setting=Offline Adaptation2025.10 | 90.2 | — | — | |
| Gemini-2.5-proSetting=Manual2026.01 | 81.1 | 109.85 | 27.65 | |
| EvoRouteSetting=Ours2026.01 | 79.5 | 65.8 | 20.53 | |
| GEPABase LLM=DeepSeek-V3.1-671B, Adaptation Setting=Offline Adaptation2025.10 | 76.4 | — | — | |
| Base LLMBase LLM=DeepSeek-V3.1-671B2025.10 | 75.2 | — | — | |
| MasRouterSetting=Routing2026.01 | 73.1 | 92.38 | 32.3 | |
| GraphRouterSetting=Routing2026.01 | 62.5 | 119.5 | 27.5 | |
| PromptLLMSetting=Routing2026.01 | 60.07 | 127.05 | 26.88 | |
| GPT-4oSetting=Manual2026.01 | 57.58 | 135.1 | 34.31 | |
| GPT-4.1Setting=Manual2026.01 | 55.3 | 89.82 | 21.95 | |
| Qwen3-14bSetting=Manual2026.01 | 41.26 | 14.16 | 31.67 |