NLI on MultiMed-X JP
78AccuracyMED-COREASONER
Evaluation Results
| Method | Links | |
|---|---|---|
| MED-COREASONERbackbone=GPT-5.12026.01 | 78 | |
| GPT-5.12026.01 | 77.33 | |
| GPT-5.22026.01 | 76.67 | |
| GPT-4o2026.01 | 76 | |
| Claude-3.5-haiku2026.01 | 68.67 |
| Method | Links | |
|---|---|---|
| MED-COREASONERbackbone=GPT-5.12026.01 | 78 | |
| GPT-5.12026.01 | 77.33 | |
| GPT-5.22026.01 | 76.67 | |
| GPT-4o2026.01 | 76 | |
| Claude-3.5-haiku2026.01 | 68.67 |