InternVL2-1B: benchmark results
Provider: Shanghai AI Lab. Released 2024-07-08. Access: Open.
Aggregate Rating 1302 ± 29, rank #8701 of 10916 rated models, from 445 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM CCBench - Calligraphy and Painting | 91.2 | Accuracy (%) | 99.3 |
| OpenVLM MMT-Bench - Image Season Recognition | 90 | Score (%) | 98.3 |
| OpenVLM MMT-Bench - Temporal Ordering | 45 | Score (%) | 97.6 |
| OpenVLM CCBench - Traditional Shows | 83.3 | Accuracy (%) | 96.5 |
| CCBench | 76.3 | Accuracy (%) | 92.5 |
| OpenVLM CCBench - Food and Clothes | 73.9 | Accuracy (%) | 92.3 |
| OpenVLM MMT-Bench - Animal Keypoint Detection | 45 | Score (%) | 92.2 |
| OpenVLM CCBench - Cultural Relics | 80.4 | Accuracy (%) | 91.9 |
| OpenVLM MMT-Bench - Geometrical Relativity | 70 | Score (%) | 87.9 |
| OpenVLM MMT-Bench - Handwritten Mathematical Expression Recognition | 90 | Score (%) | 87.4 |
| OpenVLM MMT-Bench - Video Captioning | 100 | Score (%) | 87.4 |
| OpenVLM CCBench - Scenery and Buildings | 80 | Accuracy (%) | 86.8 |
Interactive version: theaggregate.ai/model?slug=internvl2-1b · How It Works · Data refreshed daily, snapshot 2026-11-08.