LookZoom DeepEyes 98 Pixel Budget - 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. DeepEyes is separately trained Qwen2.5-VL-7B agent, official prompt/tool loop; per-image pixels B * 32^2 permit up to 128/256/512 Qwen2.5 tokens; answer-format condition separate. Five-Benchmark Average. Metric: Accuracy (%).
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | LookZoom study Qwen2.5-VL-7B, single view | 52 | |
| 2 | LookZoom study DeepEyes-7B + answer-format instruction | 42.2 | |
| 3 | LookZoom study DeepEyes-7B, official prompt | 29.8 |
Interactive version: theaggregate.ai/benchmark?slug=lookzoom-deepeyes-98-pixel-budget-five-benchmark-average · How It Works · Data refreshed daily, snapshot 2026-10-27.