Long-form QA on MultiMed-X KO
4.54Overall ScoreMED-COREASONER
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MED-COREASONERbackbone=GPT-5.12026.01 | 4.54 | 4.57 | 4.84 | 4.79 | 4.45 | 92 | |
| GPT-5.22026.01 | 4.41 | 4.46 | 4.655 | 4.815 | 4.325 | 91.5 | |
| GPT-5.12026.01 | 4.39 | 4.475 | 4.61 | 4.795 | 4.33 | 89 | |
| Claude-3.5-haiku2026.01 | 4.115 | 4.17 | 4.24 | 4.795 | 4.11 | 86 | |
| GPT-4o2026.01 | 4.055 | 4.13 | 4.105 | 4.805 | 4.225 | 84 |