HealthGPT-M3-3.8B: benchmark results
Provider: Other.
Aggregate Rating 1522 ± 52, rank #2041 of 10916 rated models, from 19 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MedQ-Deg - Treatment Reasoning | 19.2 | Accuracy (%) on the Treatment questions of MedQ-Deg (capabil | 65.4 |
| Holtercare-Bench - Diagnosis (Text) | 31.1 | Accuracy (%; multiple-choice closed QA on the overall diagno | 58.3 |
| Holtercare-Bench - Presence (Text) | 56.05 | Accuracy (%; multiple-choice closed QA on whether a given rh | 58.3 |
| MedQ-Deg - Clinical Understanding | 63.7 | Accuracy (%) on the Clinical Understanding questions of MedQ | 51.3 |
| MedQ-Deg - Diagnostic Reasoning | 27.8 | Accuracy (%) on the Diagnosis questions of MedQ-Deg (capabil | 47.4 |
| MedQ-Deg - Imaging Perception | 28.5 | Accuracy (%) on the Imaging Perception questions of MedQ-Deg | 47.4 |
| MedQ-Deg - Severe Degradation (L2) | 53.45 | Accuracy (%) on the MedQ-Deg items with severe degradations | 41 |
| MedQ-Deg - Noise Degradation | 54.7 | Accuracy (%) on MedQ-Deg items whose images carry noise degr | 38.5 |
| MedQ-Deg - Artifact Degradation | 52.3 | Accuracy (%) on MedQ-Deg items whose images carry imaging-ar | 37.2 |
| MedQ-Deg - Resolution and Blur Degradation | 55.8 | Accuracy (%) on MedQ-Deg items whose images carry resolution | 35.9 |
| MedQ-Deg - Motion Degradation | 56 | Accuracy (%) on MedQ-Deg items whose images carry motion-int | 34.6 |
| MedQ-Deg - Mild Degradation (L1) | 56.61 | Accuracy (%) on the MedQ-Deg items with mild degradations (L | 33.3 |
Interactive version: theaggregate.ai/model?slug=healthgpt-m3-3-8b · How It Works · Data refreshed daily, snapshot 2026-11-08.