DR Pixel: open-source page-native deep research agent beats OCR-based pipelines
Berkeley-NLP · hf · 2026-11-03
Researchers from Berkeley NLP, NYU Shanghai, and collaborators release DR Pixel, the first open-source end-to-end multimodal deep research agent that works directly on rendered document page images instead of OCR-extracted text, operating over a corpus of 1.06M papers.
- Motivation: scientific papers are rich in tables, figures, and equations; OCR pipelines lose layout information, while VLMs can reason page-natively.
- Benchmark: DRPIXELBENCH evaluates reasoning across text, figures, tables, and equations.
- Findings: page-native deep research already works well without training, outperforming OCR-based baselines; counterfactual credit reweighting further improves retrieval and reasoning.
Code and benchmark are open-sourced (GitHub: StarTrail-org/DRPixel; HF: HughieHu/DRPixel-Bench).
More from Research
- RaiSim 2.13.0 lets you run robotics examples online and adds a fast fluid dynamics model — ChongZzZhang · 2026-11-03
- Four low-cost coding agents survive Minecraft for 12 days, with server-log audit of every milestone — Unusual-Spot-5846 · 2026-11-03
- DAMI: dual-axis memory initialization improves long-context retrieval in gated linear attention — SpeedAI-Lab · 2026-11-03
- What OpenAI left out of its Navier-Stokes announcement — elsleightholm · 2026-11-03
- Paper rejected with two reports: the AI referee report was excellent, the human one cited themselves — birchlse · 2026-11-03
- Skyfall AI's SKY-001 world model matches GPT-6 Astra CAD accuracy at 150x lower cost — rohanpaul_ai · 2026-11-03