InternVL3-1B: benchmark results
Provider: Shanghai AI Lab. Released 2025-04-11. Access: Open.
Aggregate Rating 1376 ± 13, rank #6949 of 10916 rated models, from 617 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM MMT-Bench - Image Matting | 50 | Score (%) | 100 |
| OpenVLM MMT-Bench - 3D CAD Recognition | 85 | Score (%) | 99.8 |
| OpenVLM MMT-Bench - Scene Recognition | 85 | Score (%) | 97.3 |
| OpenVLM MMT-Bench - Helmet Anomaly Detection | 95 | Score (%) | 97.1 |
| CCBench | 78.4 | Accuracy (%) | 95.6 |
| OpenVLM MMBench V1.1 EN - Image Topic | 98.9 | Accuracy (%) | 95.2 |
| OpenVLM CCBench - Calligraphy and Painting | 87.7 | Accuracy (%) | 94.7 |
| OpenVLM POPE | 89.5 | Overall (%) | 94.2 |
| OpenVLM CCBench - Cultural Relics | 84.5 | Accuracy (%) | 94.1 |
| OpenVLM CCBench - Food and Clothes | 76.5 | Accuracy (%) | 93.8 |
| OpenVLM CCBench - Traditional Shows | 77.3 | Accuracy (%) | 93.2 |
| OpenVLM MMT-Bench - Pixel Localization | 45 | Score (%) | 93 |
Interactive version: theaggregate.ai/model?slug=internvl3-1b · How It Works · Data refreshed daily, snapshot 2026-11-08.