Medical Question Answering on MEDIQ (test)
76.4AccuracyNum. cutoff
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Num. cutoffModel=MedGemma-9B2026.06 | 76.4 | 1,412 | — | — | |
| EvolvedModel=MedGemma-9B2026.06 | 75.6 | 104 | 0.8 | 92.6 | |
| EvolvedModel=Qwen-2.5-72B2026.06 | 73.6 | 961 | 2.5 | 54.2 | |
| Num. cutoffModel=Qwen-2.5-72B2026.06 | 71.1 | 2,100 | — | — | |
| EvolvedModel=Llama-3-70B2026.06 | 62.2 | 514 | 3.6 | 67.6 | |
| Best from MEDIQModel=Llama-3-70B2026.06 | 60.9 | — | 2.3 | — | |
| BinaryModel=Llama-3-70B2026.06 | 58.6 | 1,585 | — | — | |
| EvolvedModel=Llama-3-8B2026.06 | 48.2 | 289 | 3.1 | 89.6 | |
| Best from MEDIQModel=Llama-3-8B2026.06 | 45.8 | — | 0.7 | — | |
| Num. cutoffModel=Llama-3-8B2026.06 | 45.1 | 2,788 | — | — |