Long-form QA on MultiMed-X ZH
4.53Overall ScoreMED-COREASONER
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MED-COREASONERbackbone=GPT-5.12026.01 | 4.53 | 4.55 | 4.89 | 4.78 | 4.39 | 93.5 | |
| GPT-5.12026.01 | 4.435 | 4.49 | 4.66 | 4.845 | 4.395 | 90.5 | |
| GPT-5.22026.01 | 4.42 | 4.46 | 4.605 | 4.8 | 4.36 | 91.5 | |
| GPT-4o2026.01 | 4.205 | 4.27 | 4.27 | 4.855 | 4.25 | 86 | |
| Claude-3.5-haiku2026.01 | 4.15 | 4.205 | 4.295 | 4.735 | 4.14 | 84.5 |