LLaVA-OV-0.5B: benchmark results
Provider: LLaVA Team.
Aggregate Rating 1450 ± 51, rank #3780 of 10916 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SYNCR - Multi-Angle Synchronization | 53 | Exact-match accuracy (%) on the SYNCR Multi-Angle Synchroniz | 61.8 |
| SYNCR - Object Re-identification | 31 | Exact-match accuracy (%) on the SYNCR Object Re-identificati | 38.2 |
| SYNCR - Numerical Comparison | 21 | Exact-match accuracy (%) on the SYNCR Numerical Comparison t | 32.4 |
| SYNCR - Sequential Ordering | 31 | Exact-match accuracy (%) on the SYNCR Sequential Ordering ta | 29.4 |
| SYNCR | 30.5 | Average exact-match accuracy (%) over the eight SYNCR tasks, | 23.5 |
| SYNCR - Route Planning | 15 | Exact-match accuracy (%) on the SYNCR Route Planning task (p | 23.5 |
| VisualPuzzles | 27.2 | Overall Accuracy (%) | 8.1 |
| SYNCR - Object Counting | 36.5 | Exact-match accuracy (%) on the SYNCR Object Counting task ( | 5.9 |
| TokenSwap-Bench (Image-Interleaved) | 24.7 | Accuracy (%; 1,516 MMLU questions in which visualizable conc | 2.4 |
| SYNCR - Spatial Measurement | 30.5 | Exact-match accuracy (%) on the SYNCR Spatial Measurement ta | 0 |
Interactive version: theaggregate.ai/model?slug=llava-ov-0-5b · How It Works · Data refreshed daily, snapshot 2026-11-08.