InternVL3-38B-Instruct: benchmark results
Provider: Shanghai AI Lab. Access: Open.
Aggregate Rating 1582 ± 46, rank #1223 of 10916 rated models, from 23 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MCSBench - L1 | 82.76 | Accuracy (%) | 92.1 |
| MedQ-Deg - Resolution and Blur Degradation | 67.4 | Accuracy (%) on MedQ-Deg items whose images carry resolution | 87.2 |
| MedQ-Deg - Clinical Understanding | 71.7 | Accuracy (%) on the Clinical Understanding questions of MedQ | 83.3 |
| MedQ-Deg - Severe Degradation (L2) | 62.87 | Accuracy (%) on the MedQ-Deg items with severe degradations | 82.1 |
| MCSBench - L2 | 63.66 | Accuracy (%) | 81.6 |
| MCSBench - Overall | 58.55 | Accuracy (%) | 80.3 |
| MedQ-Deg - Clean (L0) | 68.83 | Accuracy (%) on the MedQ-Deg items with the original, undegr | 79.5 |
| MedQ-Deg - Mild Degradation (L1) | 67.38 | Accuracy (%) on the MedQ-Deg items with mild degradations (L | 79.5 |
| MedQ-Deg - Intensity Degradation | 70.5 | Accuracy (%) on MedQ-Deg items whose images carry intensity- | 74.4 |
| MedQ-Deg - Noise Degradation | 64 | Accuracy (%) on MedQ-Deg items whose images carry noise degr | 74.4 |
| MedQ-Deg - Motion Degradation | 67.2 | Accuracy (%) on MedQ-Deg items whose images carry motion-int | 73.1 |
| MedQ-Deg - Anatomical Recognition | 58.3 | Accuracy (%) on the Anatomical Recognition questions of MedQ | 69.2 |
Interactive version: theaggregate.ai/model?slug=internvl3-38b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.