AeroGround - Multi-Target Relation: leaderboard
AeroGround (2026) tests vision-language models on paired UAV and ground-level views of simulated open environments. Exact-match accuracy in percent on closed-form questions; higher is better. The model orders comparison targets from nearest to farthest relative to a reference target.
Metric: Accuracy (%; Spatial Understanding: order comparison targets by distance from a reference target; aerial-ground collaborative questions from paired simulated UAV and ground views, exact-match scoring of the parsed option or mapping, invalid output scored wrong; one deterministic run). Source: arxiv.org. Saturation forecast: Around February 2027. 16 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Seed 2.0 Pro | 83.3 | #209 |
| 2 | Gemini 3.1 Pro (Preview) | 78.9 | #91 |
| 3 | GPT-5.4 | 68.9 | #121 |
| 4 | Qwen 3 VL 32B (Thinking) | 68.9 | #557 |
| 5 | Kimi K2.5 | 53.3 | #252 |
| 6 | Qwen 3 VL 8B (Thinking) | 52.8 | #887 |
| 7 | Qwen 3 VL 32B Instruct | 51.7 | #582 |
| 8 | Qwen 3 VL 8B Instruct | 49.4 | #982 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=aeroground-multi-target-relation · How It Works · Data refreshed daily, snapshot 2026-11-12.