LookZoom Pointer Comparison Two Views 196 - Five-Benchmark Average: leaderboard
LookZoom v1: Qwen3-VL-Instruct at 2B/4B/8B. Fixed 500-question subsets (seed 0) of DocVQA val (ANLS),InfographicVQA val (ANLS),ChartQA test (relaxed accuracy),TextVQA val (VQA accuracy), plus 300 held-out V-CoT-Doc (strict match). Per-view budget B means B * 32^2 pixels preserving aspect ratio, not total tokens. Single view/AttWarp/replace answer from 1 view; ViCrop/ground-then-crop/GapSight/LookZoom use 2. LookZoom reader uses GRPO LoRA; pointer trained from own correct answers; GapSight is authors reimplementation. Distinct trained systems remain separate. Appendix S12: identical adaptive window for two-view comparison; replace uses only zoomed view. Pointer and view retained in identity. Five-Benchmark Average. Metric: Accuracy (%).
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | LookZoom Qwen3-VL-8B <look> token [Two Views] | 80 | |
| 2 | LookZoom Qwen3-VL-8B pre-answer [Two Views] | 77.2 | |
| 3 | LookZoom Qwen3-VL-4B <look> token [Two Views] | 75.5 | |
| 4 | LookZoom Qwen3-VL-2B <look> token [Two Views] | 73 | |
| 5 | LookZoom Qwen3-VL-4B pre-answer [Two Views] | 72.3 | |
| 6 | LookZoom Qwen3-VL-2B pre-answer [Two Views] | 68.7 |
Interactive version: theaggregate.ai/benchmark?slug=lookzoom-pointer-comparison-two-views-196-five-benchmark-average · How It Works · Data refreshed daily, snapshot 2026-10-27.