Qwen2.5-VL Instruct (32B): benchmark results
Provider: Alibaba. Access: Open.
Aggregate Rating 1566 ± 27, rank #1460 of 10916 rated models, from 174 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HELM VHELM - Crossmodal 3600 - English Locale | 3.66 | Prometheus Vision Rating (1-5) | 97.9 |
| HELM VHELM - Flickr30K | 3.1 | Prometheus Vision rating | 97.9 |
| HELM VHELM - Math Vista - Daily Life (Multiple Choice) | 80 | Exact Match (%) | 97.9 |
| HELM VHELM - Vqa Base | 87.7 | PEM | 97.9 |
| HELM VHELM - Vqa Dialect | 85.6 | PEM (Fairness) | 97.9 |
| HELM VHELM - Crossmodal 3600 - Telugu Locale | 3.63 | Prometheus Vision Rating (1-5) | 94.7 |
| HELM VHELM - Hateful Memes | 74.18 | PEM | 93.6 |
| HELM VHELM - Toxicity | 93.62 | Mean win rate | 93.6 |
| HELM VHELM - Visual Perception | 73.58 | Mean win rate | 93.6 |
| HELM VHELM - Blink Perception - Visual Similarity | 97.78 | Prefix Quasi-Exact Match (%) | 92.6 |
| HELM VHELM - Viz Wiz | 79.3 | PEM | 91.5 |
| HELM VHELM - Crossmodal 3600 | 3.62 | Prometheus Vision rating | 87.2 |
Interactive version: theaggregate.ai/model?slug=qwen2-5-vl-instruct-32b · How It Works · Data refreshed daily, snapshot 2026-11-08.