PaliGemma (3B) Mix 224: benchmark results
Provider: Google. Released 2024-05-14. Access: API.
Aggregate Rating 1425 ± 28, rank #4930 of 10916 rated models, from 139 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HELM VHELM - Viz Wiz | 83.1 | PEM | 100 |
| HELM VHELM - Gqa | 72.2 | PEM | 97.9 |
| HELM VHELM - Blink Perception - Art Style | 100 | Prefix Quasi-Exact Match (%) | 96.8 |
| HELM VHELM - Blink Reasoning - Jigsaw | 100 | Prefix Quasi-Exact Match (%) | 96.8 |
| HELM VHELM - Blink Perception - Visual Similarity | 97.04 | Prefix Quasi-Exact Match (%) | 84 |
| HELM VHELM - Blink Perception - Relative Reflectance | 67.16 | Prefix Quasi-Exact Match (%) | 81.9 |
| HELM VHELM - Mmmu - Music | 40 | Prefix Quasi-Exact Match (%) | 73.4 |
| HELM VHELM - Vqa Robustness | 76.8 | EM (Robustness) | 71.4 |
| HELM VHELM - Vqa Base | 82.5 | PEM | 59.6 |
| HELM VHELM - Safety | 55.32 | Mean win rate | 55.3 |
| HELM VHELM - Unicorn - OODCV-VQA | 79.5 | Prefix Quasi-Exact Match (%) | 53.2 |
| HELM VHELM - Vqa Dialect | 78.3 | PEM (Fairness) | 51.1 |
Interactive version: theaggregate.ai/model?slug=paligemma-3b-mix-224 · How It Works · Data refreshed daily, snapshot 2026-11-08.