llava-v1.6-mistral-7B: benchmark results
Provider: LLaVA Team. Access: Open.
Aggregate Rating 1460 ± 25, rank #3304 of 10916 rated models, from 72 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Mind's Eye - Dynamic Structural Correspondence | 32.4 | Accuracy (%) on the dynamic structural correspondence task o | 100 |
| Mind's Eye - Mental Transformation | 35.8 | Accuracy (%) on the mental transformation (polycube rotation | 100 |
| WildFireVQA (Thermal + Radiometric Statistics) | 41.23 | Accuracy (%) over all WildFireVQA wildfire-monitoring multip | 100 |
| WildFireVQA (Thermal) | 38.74 | Accuracy (%) over all WildFireVQA wildfire-monitoring multip | 100 |
| WildFireVQA - Classification | 49.56 | Accuracy (%) on the 6 classification question types of the W | 100 |
| WildFireVQA - Distribution and Segmentation | 48.78 | Accuracy (%) on the 8 distribution and segmentation question | 100 |
| WildFireVQA - Localization and Direction | 61.5 | Accuracy (%) on the 4 localization and direction question ty | 100 |
| WildFireVQA - Presence and Detection | 78.01 | Accuracy (%) on the 8 presence and detection question types | 100 |
| Mind's Eye - Visual Conceptual Slippage | 30.6 | Accuracy (%) on the visual conceptual slippage (odd one out | 88.2 |
| GeoHeight-Bench - Threshold Segmentation | 44.99 | Text accuracy (%) on the Threshold Segmentation task (the mo | 75 |
| SaLAD - Safe Queries | 99.68 | Accuracy rate (%; GPT-4o judge; 943 benign queries answered | 70.6 |
| WildFireVQA | 52.68 | Accuracy (%) over all WildFireVQA wildfire-monitoring multip | 66.7 |
Interactive version: theaggregate.ai/model?slug=llava-v1-6-mistral-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.