LLaVA-v1-7B: benchmark results
Provider: LLaVA Team. Released 2023-04-17.
Aggregate Rating 1231 ± 50, rank #9224 of 10489 rated models, from 399 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM MMT-Bench - Font Recognition | 35 | Score (%) | 88.1 |
| OpenVLM MMT-Bench - Deepfake Detection | 60 | Score (%) | 83 |
| OpenVLM COCO Captions | 5.5 | CIDEr | 74.8 |
| OpenVLM MMT-Bench - GUI Install | 45 | Score (%) | 71.6 |
| OpenVLM MMMU - Energy and Power | 46.7 | Accuracy (%) | 71 |
| OpenVLM MMT-Bench - Action Quality Assessment | 25 | Score (%) | 69.7 |
| OpenVLM MMT-Bench - Vehicle Recognition | 100 | Score (%) | 69.4 |
| OpenVLM MMMU - Biology | 46.7 | Accuracy (%) | 69.1 |
| OpenVLM MMT-Bench - Handwritten Retrieval | 30 | Score (%) | 67.2 |
| OpenVLM MMT-Bench - Image Quality Assessment | 45 | Score (%) | 66.5 |
| OpenVLM MMT-Bench - Google Apps | 35 | Score (%) | 61.2 |
| OpenVLM MMT-Bench - Chemical Apparatus Recognition | 60 | Score (%) | 57.8 |
Interactive version: theaggregate.ai/model?slug=llava-v1-7b · How It Works · Data refreshed daily, snapshot 2026-11-20.