VeriOCRBench - Visual Dimension (Assisted): leaderboard

Metric: Diagnosis accuracy (%; the 400 trap-injected tasks of the visual verification dimension (image degradation and occluded target: the required text or visual evidence is unreadable or hidden); the model must reject the task and name the ground-truth trap mechanism; judged by GPT-4o at temperature 0; assisted paradigm: a meta-instruction asks the model to verify the task first). Source: arxiv.org. Saturation forecast: Around July 2028. 15 models tracked.

Top models

#ModelScore
1Qwen 3.5 397B A17B72
2Qwen 3.5 122B A10B71
3Claude Sonnet 4.570.75
4GPT-569.25
5Qwen 3.5 35B A3B68.5
6GPT-4o66.25
7InternVL3-38B65.75
8Llama 4 Scout56.5
9Llama 4 Maverick55.25
10Gemini 3.1 Pro (Preview)12.5

Interactive version: theaggregate.ai/benchmark?slug=veriocrbench-visual-dimension-assisted · How It Works · Data refreshed daily, snapshot 2026-09-26.