SpatialUAV - Aerial-Aerial Shared Association: leaderboard
Cross-view correspondence between two UAV views from SpatialUAV (Zhang et al., 2026): the model lists which labelled regions in the two views show the same objects, and each answer is scored by pair-level F1 against the annotated region pairs. Mean F1 in percent over 75 questions; higher is better.
Metric: Shared Association Pair F1 (%; region-pair correspondences between two UAV views). Source: arxiv.org. Saturation forecast: Around September 2027. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 | 32.8 | #121 |
| 2 | Claude Opus 4.7 | 26.2 | #75 |
| 3 | Qwen 3.6 27B | 24.7 | #339 |
| 4 | Qwen 3.6 35B A3B | 16.8 | #461 |
| 5 | Qwen 3.5 9B | 11.1 | #803 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=spatialuav-aerial-aerial-shared-association · How It Works · Data refreshed daily, snapshot 2026-11-12.