Natural Language Inference on MultiMed-X ZH
0.7667AccuracyMED-COREASONER
Evaluation Results
| Method | Links | |
|---|---|---|
| MED-COREASONERbackbone=GPT-5.12026.01 | 0.7667 | |
| GPT-5.12026.01 | 0.7467 | |
| GPT-4o2026.01 | 0.7467 | |
| GPT-5.22026.01 | 0.74 | |
| Claude-3.5-haiku2026.01 | 0.6933 |
| Method | Links | |
|---|---|---|
| MED-COREASONERbackbone=GPT-5.12026.01 | 0.7667 | |
| GPT-5.12026.01 | 0.7467 | |
| GPT-4o2026.01 | 0.7467 | |
| GPT-5.22026.01 | 0.74 | |
| Claude-3.5-haiku2026.01 | 0.6933 |