Long-form QA on MultiMed-X EN
4.6Overall ScoreMED-COREASONER
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MED-COREASONERbackbone=GPT-5.12026.01 | 4.6 | 4.64 | 4.85 | 4.86 | 4.46 | 93.9 | |
| GPT-5.12026.01 | 4.425 | 4.485 | 4.59 | 4.88 | 4.385 | 91.5 | |
| GPT-5.22026.01 | 4.415 | 4.455 | 4.615 | 4.815 | 4.41 | 90 | |
| GPT-4o2026.01 | 4.21 | 4.27 | 4.245 | 4.845 | 4.38 | 85.5 | |
| Claude-3.5-haiku2026.01 | 4.2 | 4.24 | 4.265 | 4.83 | 4.305 | 87 |