llava-v1.6-mistral-7B-hf: benchmark results
Provider: LLaVA Team. Access: Open.
Aggregate Rating 1406 ± 31, rank #5775 of 10916 rated models, from 61 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LogiScope-VQA - Equipment Operation Compliance | 82 | Micro-averaged accuracy (%) | 100 |
| LogiScope-VQA - Perimeter Access Control | 85 | Micro-averaged accuracy (%) | 100 |
| LogiScope-VQA - Personnel Safety Duty | 69 | Micro-averaged accuracy (%) | 92.1 |
| Drive-P2D - Weather and Lighting (Scene-1) | 87.15 | Accuracy (%; recognize weather and illumination conditions ( | 70 |
| VRIQ (Abstract) - Odd One Out | 27 | Accuracy (%) on the 200 abstract-domain VRIQ odd-one-out puz | 58.3 |
| MERA Multi - ruNaturalScienceVQA | 16.7 | Score (%) | 52.5 |
| MERA Multi - ruHHH-Image | 10.5 | Score (%) | 50.8 |
| Drive-P2D - Special Scene Factors (Scene-2) | 32.49 | Accuracy (%; identify special scene factors such as roadwork | 50 |
| MERA Multi - LabTabVQA | 6.5 | Score (%) | 50 |
| VRIQ (Abstract) - 3D Visualization | 28.5 | Accuracy (%) on the 200 abstract-domain VRIQ 3D visualizatio | 50 |
| LogiScope-VQA - Spatial Relation | 21 | Micro-averaged accuracy (%) | 47.4 |
| MERA Multi - ruTiE-Image | 24.8 | Score (%) | 46.6 |
Interactive version: theaggregate.ai/model?slug=llava-v1-6-mistral-7b-hf · How It Works · Data refreshed daily, snapshot 2026-11-08.