VisualTextTrap (LLaVA-Video): leaderboard
Metric: Overall VQA accuracy (%) on the VisualTextTrap videos built on LLaVA-Video (text overlays rendered on every frame under text-free, text-congruent and text-contradictory conditions, the contradictory overlays written by Claude-Sonnet-4.6), multiple-choice video QA under the default configuration of each model; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 73.7 |
| 2 | Qwen 3 VL 235B A22B | 55.3 |
| 3 | Qwen 3 VL 8B Instruct | 29.3 |
Interactive version: theaggregate.ai/benchmark?slug=visualtexttrap-llava-video · How It Works · Data refreshed daily, snapshot 2026-10-07.