Embodied-BenchClaw Aerial-Bench - Single Choice: leaderboard

Metric: Accuracy (%) on single-choice selection questions; questions over aerial and UAV images from the 5,000-question benchmark built by the Embodied-BenchClaw pipeline (object and category recognition, counting with interval choices, image-plane spatial relations, visible-area and scale comparison, depth-based near-far reasoning), vision-language setting; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 14 models tracked.

Top models

#ModelScore
1Gemini 3 Pro (Preview)90
2GPT-5.589
3Kimi K2.583
4Claude Opus 4.782
5Claude Sonnet 4.679
6Grok 4.20 0309 (Reasoning)79
7Qwen 3.6 35B A3B78
8Qwen 3.6 27B77
9GPT-5.274
10GPT-4.168
11Claude Haiku 4.5 (20251001)55

Interactive version: theaggregate.ai/benchmark?slug=embodied-benchclaw-aerial-bench-single-choice · How It Works · Data refreshed daily, snapshot 2026-09-29.