VideoVIBE - Visual Hierarchy: leaderboard

Metric: Failure-diagnosis accuracy (%; questions of the Visual Hierarchy subdimension of Visual and Motion Fidelity; zero-shot single-choice diagnosis from the interaction video and the page source code (Video+HTML), about 1.7K questions built from 6,338 verified failures of one-shot generated web applications recorded under human operation; free-form answers mapped deterministically to an option). Source: arxiv.org. Saturation forecast: Around 2034. 13 models tracked.

Top models

#ModelScore
1Qwen 3 VL 8B (Thinking)51.11
2Gemini 2.5 Flash43.11
3Gemini 3 Flash40
4Gemini 3.1 Pro (Preview)39.11
5Qwen 3 VL 8B Instruct39.11
6Qwen 3 VL 30B A3B (Thinking)39.11
7Qwen 3.5 35B A3B36.89
8GLM-4.6V36.89
9MiMo-V2.536.44
10Qwen 3 VL 30B A3B Instruct31.11
11Gemini 3.1 Flash Lite30.67

Interactive version: theaggregate.ai/benchmark?slug=videovibe-visual-hierarchy · How It Works · Data refreshed daily, snapshot 2026-09-29.