LinkS2Bench - Target Visibility: leaderboard

Metric: Target Visibility: yes/no accuracy (%) on whether a marked satellite region appears in the drone video on the full LinkS2Bench (17,903 questions over 1,500 UAV videos paired with satellite images), temperature 0; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 18 models tracked.

Top models

#ModelScore
1Gemini 3.1 Pro (Preview)84.4
2GPT-5.481.4
3Llama 4 Scout79.6
4Seed 1.876.7
5GPT-5.4 Mini76.3
6Qwen 3.5 35B A3B71.4
7Qwen 3.5 397B A17B66.7
8Kimi K2.555.1
9Claude Sonnet 4.653.3
10Qwen 3.5 4B27
11Qwen 3.5 9B26.3
12Ministral 3 3B17.7
13Ministral 3 8B15.3
14Ministral 3 14B13.9

Interactive version: theaggregate.ai/benchmark?slug=links2bench-target-visibility · How It Works · Data refreshed daily, snapshot 2026-10-07.