Leaderboard
Health LLM Benchmark Scores
All scores reproducible: temperature 0, seed 42, 100 samples per eval. Submit your model →
| # | Model | Med Knowledge | Clinical Reasoning | Evidence | Claims | NER | JSON | Consistency | Health Score | Grade |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | qwen2.5:7bLocal machine | 51.0 | 67.0 | 30.0 | 23.0 | 22.4 | 27.0 | 68.0 | 41.2 | D |