DisasterBench (UAV): leaderboard
Metric: Overall accuracy (%) over the nine response tasks on the 3,000-question test split of low-altitude UAV disaster images (14 disaster scene types), multiple choice scored by strict exact match on the final option letter; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 21 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 84.83 |
| 2 | Gemini 2.5 Pro | 84.17 |
| 3 | Claude Sonnet 4.5 | 79.93 |
| 4 | Gemini 2.0 Flash | 78.8 |
| 5 | Gemini 2.5 Flash | 78.03 |
| 6 | Grok 4 | 77.8 |
| 7 | GPT-4o | 73 |
| 8 | Qwen 2.5 VL 7B Instruct | 71.6 |
| 9 | InternVL2-8B | 68.23 |
| 10 | InternVL3.5-8B | 66.4 |
| 11 | GPT-4o Mini | 64 |
Interactive version: theaggregate.ai/benchmark?slug=disasterbench-uav · How It Works · Data refreshed daily, snapshot 2026-09-29.