WildRoadBench - Road Crack: leaderboard
The road crack class of WildRoadBench (2026), an aerial road-damage grounding benchmark: models box every instance in wild drone images, 561 ground-truth boxes in all. The score is F1 at IoU 0.5 in percent; higher is better.
Metric: F1 (%) of predicted boxes at IoU 0.5 for the road crack scene class (561 ground-truth boxes), VLM Track, zero-shot box grounding on the full 1,061-image aerial set (1,394 ground-truth boxes); higher is better. Source: arxiv.org. Saturation forecast: Around October 2027. 25 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro | 33.3 | #146 |
| 2 | Qwen 3.6 Plus | 25.4 | #245 |
| 3 | Gemini 3.1 Flash Lite | 16.2 | #377 |
| 4 | Qwen 2.5 VL 7B Instruct | 15.3 | #1919 |
| 5 | Qwen2.5-VL-32B-Instruct-AWQ | 11.7 | |
| 6 | Claude Sonnet 4.6 | 11.3 | #179 |
| 7 | GPT-5.4 (2026-03-05) | 10.8 | #95 |
| 8 | Qwen3.5-VL-27B-FP8 | 7.4 | |
| 9 | Qwen3.5-VL-35B-A3B-FP8 | 7.3 | |
| 10 | Qwen3-VL-32B-Instruct-AWQ | 6.5 | |
| 11 | Qwen 3 VL 8B Instruct | 6.2 | #1043 |
| 12 | Qwen 3.5 VL 9B | 5.6 | |
| 13 | Qwen3-VL-32B-Thinking-AWQ | 5.1 | |
| 14 | Qwen 3.5 VL 4B | 4.3 | |
| 15 | Kimi K2.5 | 4 | #272 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=wildroadbench-road-crack · How It Works · Data refreshed daily, snapshot 2026-11-22.