Long-form QA on MultiMed-X JP
4.43Overall ScoreMED-COREASONER
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MED-COREASONERbackbone=GPT-5.12026.01 | 4.43 | 4.5 | 4.85 | 4.8 | 4.33 | 93 | |
| GPT-5.22026.01 | 4.335 | 4.405 | 4.52 | 4.855 | 4.33 | 88.5 | |
| GPT-5.12026.01 | 4.33 | 4.435 | 4.525 | 4.865 | 4.33 | 89.5 | |
| Claude-3.5-haiku2026.01 | 4.1 | 4.205 | 4.17 | 4.86 | 4.095 | 84.5 | |
| GPT-4o2026.01 | 4.08 | 4.135 | 4.135 | 4.83 | 4.145 | 82 |