Llama3-Med42-8B: benchmark results
Provider: Meta.
Aggregate Rating 1425 ± 29, rank #4901 of 10916 rated models, from 38 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MHGraphBench - Two-hop Verification | 55 | Accuracy (%) on the 1,200 two-hop verification items (about | 85.7 |
| MedPriv-Bench - Fact-Level Leakage | 72.8 | Fact-level leakage rate (%): share of all injected sensitive | 85.7 |
| MedPriv-Bench - Instance-Level Leakage | 85.5 | Instance-level leakage rate (%): share of answers that revea | 78.6 |
| MedPriv-Bench (Privacy Prompt) - Fact-Level Leakage | 67.6 | Fact-level leakage rate (%): share of all injected sensitive | 57.1 |
| ShatterMed-QA (English, Easy) | 82.64 | Accuracy (%) on the 5,923 English Easy multiple-choice clini | 55 |
| SurgiQ - Negative | 56.27 | Negative (all-except) questions (2,153): accuracy (%) on Sur | 52.9 |
| MEDIC Benchmark | 62.84 | MEDIC Public Table Average | 52.6 |
| MHGraphBench - Relation Prediction | 27.97 | Accuracy (%) on the 1,634 four-way drug-disease relation pre | 50 |
| ShatterMed-QA (English, Hard) | 80.45 | Accuracy (%) on the 1,692 English Hard multiple-choice clini | 50 |
| SurgiQ - Critical Care and Emergency | 55.25 | Critical care and emergency surgery questions (590): accurac | 45.6 |
| SurgiQ - General Surgery | 56.48 | General surgery questions (6,948): accuracy (%) on SurgiQ's | 44.1 |
| MHGraphBench - Entity Clustering | 24.55 | Accuracy (%) on the 2,000 five-entity odd-one-out entity clu | 42.9 |
Interactive version: theaggregate.ai/model?slug=llama3-med42-8b · How It Works · Data refreshed daily, snapshot 2026-11-08.