LLaVA-OV-7B: benchmark results
Provider: LLaVA Team.
Aggregate Rating 1442 ± 35, rank #4139 of 10916 rated models, from 35 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SYNCR - Multi-Angle Synchronization | 56 | Exact-match accuracy (%) on the SYNCR Multi-Angle Synchroniz | 82.4 |
| SYNCR - Object Re-identification | 37.5 | Exact-match accuracy (%) on the SYNCR Object Re-identificati | 70.6 |
| SYNCR - Route Planning | 21.5 | Exact-match accuracy (%) on the SYNCR Route Planning task (p | 52.9 |
| JMMMU-Pro - Japanese Art | 30.67 | Accuracy (%) | 46.2 |
| Molmo2-ER Release - MindCube | 45.6 | Benchmark score (%) | 45.5 |
| Molmo2-ER Release - RoboSpatial-VQ | 68.9 | Benchmark score (%) | 45.5 |
| SYNCR - Numerical Comparison | 22.5 | Exact-match accuracy (%) on the SYNCR Numerical Comparison t | 41.2 |
| SYNCR - Sequential Ordering | 32 | Exact-match accuracy (%) on the SYNCR Sequential Ordering ta | 41.2 |
| JMMMU-Pro - Japanese Heritage | 26.67 | Accuracy (%) | 38.5 |
| SYNCR | 33.5 | Average exact-match accuracy (%) over the eight SYNCR tasks, | 35.3 |
| BareBones - WTP-Bench (Silhouette) | 1.81 | Exact-match accuracy (%) naming the Pokemon from its pure-sh | 27.3 |
| SpatialBabel | 64 | Three.js (self-reported) | 26.9 |
Interactive version: theaggregate.ai/model?slug=llava-ov-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.