NVILA-8B: benchmark results
Provider: Other. Access: Open.
Aggregate Rating 1523 ± 32, rank #2029 of 10916 rated models, from 42 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| GeoMMBench - Multispectral and Hyperspectral | 67.6 | Micro-averaged accuracy (%) on the multispectral and hypersp | 80.9 |
| GST-Bench - Global Position (Semantic) | 26.64 | Mean accuracy (%; predicted top-down map position of a previ | 76.2 |
| XNote (Naive RAG) | 70.4 | Naive-RAG deception-detection accuracy (%) on all 2,176 XNot | 75 |
| PhysBench | 43.8 | Overall Accuracy (%) | 70 |
| XNote | 78.45 | Closed-book deception-detection accuracy (%) on all 2,176 XN | 66.7 |
| GST-Bench - Egocentric Direction (Visual) | 18.15 | Mean accuracy (%; share of predicted egocentric directions t | 57.1 |
| GeoMMBench - Photogrammetry | 57.5 | Micro-averaged accuracy (%) on the photogrammetry discipline | 52.9 |
| GST-Bench - Egocentric Distance (Visual) | 7.22 | Mean relative accuracy (%; distance to a previously seen obj | 52.4 |
| GST-Bench - Top-Down Selection (Hard) | 29.78 | Accuracy (%; multiple choice of the top-down image of the ex | 52.4 |
| GST-Bench - Trajectory Selection | 25.59 | Accuracy (%; multiple choice of the camera trajectory drawn | 50 |
| GeoMMBench - DEM | 45.5 | Micro-averaged accuracy (%) on the digital elevation model q | 50 |
| GeoMMBench - Applications | 51.7 | Micro-averaged accuracy (%) on the remote sensing applicatio | 48.5 |
Interactive version: theaggregate.ai/model?slug=nvila-8b · How It Works · Data refreshed daily, snapshot 2026-11-08.