PaveVQA - Classification Accuracy: leaderboard
PaveVQA is the vision-language question answering part of PaveBench (2026), built on real highway pavement inspection images with single-turn, multi-turn and expert-corrected questions about distress recognition, localization, quantification and maintenance. This board is accuracy in percent on categorical questions about distress presence and type, answered zero-shot; higher is better.
Metric: Classification accuracy (%) on categorical questions about distress presence and type, of the zero-shot model on PaveVQA, the vision-language question answering part of PaveBench (real highway pavement images, single-turn, multi-turn and expert-corrected questions); higher is better. Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 3 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 VL 3B | 65.18 | #2639 |
| 2 | LLaVA-OneVision-7B | 60.91 | #2981 |
| 3 | DeepSeek-VL2-Small | 55.48 | #2464 |
Interactive version: theaggregate.ai/benchmark?slug=pavevqa-classification-accuracy · How It Works · Data refreshed daily, snapshot 2026-11-21.