Meditron-7B: benchmark results
Provider: EPFL.
Aggregate Rating 1290 ± 35, rank #8864 of 10916 rated models, from 32 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| ShatterMed-QA (Chinese, Easy) | 92.34 | Accuracy (%) on the 2,616 Chinese Easy multiple-choice clini | 55 |
| ShatterMed-QA (English, Easy) | 80.6 | Accuracy (%) on the 5,923 English Easy multiple-choice clini | 45 |
| ClinicBench | 36.32 | Average Score | 38.1 |
| MHGraphBench - Entity Typing | 20.25 | Accuracy (%) on the 1,847 five-way entity typing items, lett | 28.6 |
| MHGraphBench - Fact Checking | 50 | Accuracy (%) on the 4,000 balanced yes/no triple fact checki | 28.6 |
| MHGraphBench - Relation Prediction | 25.34 | Accuracy (%) on the 1,634 four-way drug-disease relation pre | 28.6 |
| ShatterMed-QA (Chinese, Hard) | 73.09 | Accuracy (%) on the 327 Chinese Hard multiple-choice clinica | 25 |
| MHGraphBench - Two-hop Verification with Evidence | 51 | Accuracy (%) on the 1,200 two-hop verification items with sa | 21.4 |
| ShatterMed-QA (English, Hard) | 64.74 | Accuracy (%) on the 1,692 English Hard multiple-choice clini | 15 |
| MHGraphBench - Entity Clustering | 18 | Accuracy (%) on the 2,000 five-entity odd-one-out entity clu | 14.3 |
| SurgiQ - Critical Care and Emergency | 28.64 | Critical care and emergency surgery questions (590): accurac | 8.8 |
| CureMed-Bench | 2.5 | Logical accuracy (%; share of final answers a GPT-4o judge m | 6.2 |
Interactive version: theaggregate.ai/model?slug=meditron-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.