InternVL2.5-38B: benchmark results
Provider: Shanghai AI Lab. Released 2024-12-05. Access: Open.
Aggregate Rating 1564 ± 12, rank #1492 of 10916 rated models, from 311 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Medical MM Leaderboard - PMC-VQA | 79.9 | Accuracy (%) | 100 |
| OpenVLM AI2D - Rock Cycle | 80.6 | Accuracy (%) | 98.8 |
| OpenVLM MMBench V1.1 EN - Image Scene | 92.4 | Accuracy (%) | 98.4 |
| OpenVLM MMBench V1.1 CN - Physical Relation | 74.7 | Accuracy (%) | 98.1 |
| OpenVLM AI2D - Photosynthesis and Respiration | 89.9 | Accuracy (%) | 97.9 |
| OpenVLM MMBench V1.1 CN - Image Style | 84.8 | Accuracy (%) | 97.5 |
| PhysBench | 51.9 | Overall Accuracy (%) | 97.5 |
| OpenVLM MMBench V1.1 EN - Physical Relation | 74.7 | Accuracy (%) | 96.8 |
| OpenVLM MMBench V1.1 CN - Image Scene | 92.4 | Accuracy (%) | 96.6 |
| OpenVLM AI2D - Atom Structure | 100 | Accuracy (%) | 96.5 |
| OpenVLM AI2D - Faults and Earthquakes | 85.7 | Accuracy (%) | 96.5 |
| OpenVLM AI2D - Rock Strata | 90.2 | Accuracy (%) | 96.5 |
Interactive version: theaggregate.ai/model?slug=internvl2-5-38b · How It Works · Data refreshed daily, snapshot 2026-11-08.