LLaVA-1.5-13B: benchmark results
Provider: LLaVA Team. Released 2023-10-05. Access: Open.
Aggregate Rating 1418 ± 28, rank #5268 of 10916 rated models, from 116 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| PuzzleMate - Piece Orientation Verification (GoPro) | 93 | Accuracy (%; exact answer; task: binary judgment of whether | 92.9 |
| EIBench | 54.37 | Recall | 91.7 |
| GQA | 80.34 | Overall Accuracy (%) | 88.7 |
| EIBench (complex) | 38.1 | Recall | 75 |
| REBUS | 1.8 | Accuracy | 71.4 |
| PathBind - Text-only VQA D3 | 34 | Score (%) | 67.6 |
| PathBind - Text-only PathBind Average | 36.5 | Score (%) | 64.7 |
| PathBind - Text-only VQA D6 | 26.4 | Score (%) | 64.7 |
| PlantMicro - Localization mAP@0.5 | 30.78 | Mean average precision at IoU 0.5 (%) of directly predicted | 62.5 |
| PathBind - Text-only PTA D1 | 37.5 | Score (%) | 58.8 |
| PathBind - Text-only Quilt-VQA YorN | 44.2 | Score (%) | 58.8 |
| PathBind - Text-only VQA D2 | 34.4 | Score (%) | 58.8 |
Interactive version: theaggregate.ai/model?slug=llava-1-5-13b · How It Works · Data refreshed daily, snapshot 2026-11-08.