DR Pixel: Page-Native Deep Research with Vision-Language Models
- Published
- Source
- arXiv
- Paper number
- 1402
- Field
- Research Agents
- arXiv ID
- 2611.03097
Key points
- DR Pixel is the first open-source end-to-end multimodal deep research agent that uses rendered page images as its sole retrieval and reasoning unit over a corpus of 1.06M pages.
- The authors built DRPixelBench with 500 questions on scientific papers, where every question requires reading text, tables, figures, and equations together.
- Without training, the page-native approach improved accuracy by 5.1–8.4 points over OCR-based pipelines across GPT-5.4, Gemini-3.5-Flash, and Qwen3.6-27B while using 20%–30% fewer retrieval calls.
- Qwen3.6-27B trained with counterfactual credit reweighting raised DRPixelBench accuracy from 40.1 to 52.1, slightly surpassing GPT-5.4.
- The training came within 1.8 points of an oracle on out-of-distribution ViDoRe-V3 and needed 57% fewer searches than vanilla SFT for a comparable hit rate.
Paper links
External research summaries. These are not HDATF publications or measured product results.