InternVL2-2B: benchmark results
Provider: Shanghai AI Lab. Released 2024-07-04. Access: Open.
Aggregate Rating 1360 ± 19, rank #7437 of 10916 rated models, from 1034 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM MMT-Bench - Animal Keypoint Detection | 65 | Score (%) | 100 |
| MEGA-Bench Task - Hashtag Recommendation | 97.6 | Task Score (%) | 98.8 |
| OpenVLM CCBench - Food and Clothes | 78.3 | Accuracy (%) | 95.6 |
| OpenVLM CCBench - Sketch Reasoning | 93.3 | Accuracy (%) | 95.2 |
| OpenVLM CCBench - Cultural Relics | 84.5 | Accuracy (%) | 94.1 |
| MEGA-Bench Task - ASCII Art 30 | 35.7 | Task Score (%) | 93 |
| OpenVLM MMT-Bench - Gaze Estimation | 35 | Score (%) | 93 |
| OpenVLM MMT-Bench - Crowd Counting | 70 | Score (%) | 92.7 |
| OpenVLM MMT-Bench - Temporal Ordering | 40 | Score (%) | 92 |
| MEGA-Bench Task - Face Keypoint Detection | 84.9 | Task Score (%) | 91.9 |
| OpenVLM SEEDBench IMG - Text Understanding | 75 | Accuracy (%) | 91.6 |
| OpenVLM MMT-Bench - 3D Indoor Recognition | 45 | Score (%) | 91.5 |
Interactive version: theaggregate.ai/model?slug=internvl2-2b · How It Works · Data refreshed daily, snapshot 2026-11-08.