Clinical Diagnosis on RareBench (combined)
39.31Recall@1Mixed-Vendor MAC
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Mixed-Vendor MACSetup=Mixed-Vendor MAC2026.02 | 39.31 | 49.82 | 55.05 | 61.35 | |
| Single-LLM (Gemini)Setup=Single-LLM, Model=Gemini-2.5-Pro2026.02 | 37.58 | 47.79 | 50.84 | 56.41 | |
| Single-Vendor MAC (Gemini)Setup=Single-Vendor MAC, Model=Gemini-2.5-Pro2026.02 | 36.63 | 43.43 | 46.47 | 51.93 | |
| Single-Vendor MAC (Claude)Setup=Single-Vendor MAC, Model=Claude-4.5-Sonnet2026.02 | 33.32 | 48.06 | 50.53 | 57.06 | |
| Single-Vendor MAC (OpenAI)Setup=Single-Vendor MAC, Model=o4-mini2026.02 | 32.73 | 41.22 | 45.54 | 52.33 | |
| Single-LLM (OpenAI)Setup=Single-LLM, Model=o4-mini2026.02 | 31.81 | 40.52 | 46.12 | 53.07 | |
| Single-LLM (Claude)Setup=Single-LLM, Model=Claude-4.5-Sonnet2026.02 | 29.35 | 39.54 | 45.41 | 55.42 |