InternVL3-78B-Instruct: benchmark results
Provider: Shanghai AI Lab. Access: Open.
Aggregate Rating 1562 ± 34, rank #1518 of 10916 rated models, from 27 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MedQ-Deg - Clean (L0) | 74.04 | Accuracy (%) on the MedQ-Deg items with the original, undegr | 100 |
| MedQ-Deg - Clinical Understanding | 75.9 | Accuracy (%) on the Clinical Understanding questions of MedQ | 100 |
| MedQ-Deg - Mild Degradation (L1) | 72.46 | Accuracy (%) on the MedQ-Deg items with mild degradations (L | 100 |
| MedQ-Deg - Motion Degradation | 75.6 | Accuracy (%) on MedQ-Deg items whose images carry motion-int | 100 |
| MedQ-Deg - Resolution and Blur Degradation | 72.7 | Accuracy (%) on MedQ-Deg items whose images carry resolution | 100 |
| MedQ-Deg - Severe Degradation (L2) | 68.63 | Accuracy (%) on the MedQ-Deg items with severe degradations | 100 |
| MedQ-Deg - Artifact Degradation | 64.3 | Accuracy (%) on MedQ-Deg items whose images carry imaging-ar | 97.4 |
| MedQ-Deg - Noise Degradation | 69.3 | Accuracy (%) on MedQ-Deg items whose images carry noise degr | 97.4 |
| MedQ-Deg - Intensity Degradation | 75.1 | Accuracy (%) on MedQ-Deg items whose images carry intensity- | 92.3 |
| MCSBench - L1 | 80.46 | Accuracy (%) | 90.1 |
| MedQ-Deg - Anatomical Recognition | 66.3 | Accuracy (%) on the Anatomical Recognition questions of MedQ | 89.7 |
| MedQ-Deg - Imaging Perception | 35 | Accuracy (%) on the Imaging Perception questions of MedQ-Deg | 89.7 |
Interactive version: theaggregate.ai/model?slug=internvl3-78b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.