mlsanityData to 7 Oct 2026

DeepSeek-OCR 2: Visual Causal Flow

arXiv 2601.20552 ↗First posted 28 Jan 2026cs.CVcited by 68

Not on the map. Why?

Abstract (arXiv)

We present DeepSeek-OCR 2 to investigate the feasibility of a novel encoder-DeepEncoder V2-capable of dynamically reordering visual tokens upon image semantics. Conventional vision-language models (VLMs) invariably process visual tokens in a rigid raster-scan order (top-left to bottom-right) with fixed positional encoding when fed into LLMs. However, this contradicts human visual perception, which follows flexible yet semantically coherent scanning patterns driven by inherent logical structures. Particularly for images with complex layouts, human vision exhibits causally-informed sequential processing. Inspired by this cognitive mechanism, DeepEncoder V2 is designed to endow the encoder with causal reasoning capabilities, enabling it to intelligently reorder visual tokens prior to LLM-based content interpretation. This work explores a novel paradigm: whether 2D image understanding can be effectively achieved through two-cascaded 1D causal reasoning structures, thereby offering a new architectural approach with the potential to achieve genuine 2D reasoning. Codes and model weights are publicly accessible at http://github.com/deepseek-ai/DeepSeek-OCR-2.

Key facts. Cited by 68 papers in the map (data through 7 Oct 2026). Builds on 41 of its 48 references that are in the map. First posted on arXiv 28 Jan 2026.

Builds on

41 of 48 references.

References · 41

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  3. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  4. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  5. 2015Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networkscited by 9,617
  6. 2023Segment Anythingcited by 8,158
  7. 2020End-to-End Object Detection with Transformerscited by 7,384
  8. 2023BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Modelscited by 6,069
  9. 2025Qwen2.5-VL Technical Reportcited by 5,418
  10. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,581
  11. 2024Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolutioncited by 4,151
  12. 2020Deformable DETR: Deformable Transformers for End-to-End Object Detectioncited by 3,410
29 more
  1. 2025Qwen3-VL Technical Reportcited by 2,534
  2. 2023Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyondcited by 2,179
  3. 2025InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Modelscited by 1,769
  4. 2016YOLO9000: Better, Faster, Strongercited by 1,429
  5. 2025InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiencycited by 1,419
  6. 2024DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Modelcited by 1,093
  7. 2020Multilingual Denoising Pre-training for Neural Machine Translationcited by 1,028
  8. 2024Chameleon: Mixed-Modal Early-Fusion Foundation Modelscited by 904
  9. 2025DeepSeek-V3.2: Pushing the Frontier of Open Large Language Modelscited by 755
  10. 2023Neural Codec Language Models are Zero-Shot Text to Speech Synthesizerscited by 737
  11. 2024MinerU: An Open-Source Solution for Precise Document Content Extractioncited by 255
  12. 2022OPT-IML: Scaling Language Model Instruction Meta Learning through the Lens of Generalizationcited by 235
  13. 2023Patch n' Pack: NaViT, a Vision Transformer for any Aspect Ratio and Resolutioncited by 226
  14. 2024General OCR Theory: Towards OCR-2.0 via a Unified End-to-end Modelcited by 174
  15. 2025DeepSeek-OCR: Contexts Optical Compressioncited by 174
  16. 2024OmniDocBench: Benchmarking Diverse PDF Document Parsing with Comprehensive Annotationscited by 144
  17. 2025olmOCR: Unlocking Trillions of Tokens in PDFs with Vision Language Modelscited by 120
  18. 2025PaddleOCR-VL: Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language Modelcited by 84
  19. 2025MonkeyOCR: Document Parsing with a Structure-Recognition-Relation Triplet Paradigmcited by 81
  20. 2025Dolphin: Document Image Parsing via Heterogeneous Anchor Promptingcited by 78
  21. 2024Focus Anywhere for Fine-grained Multi-page Document Understandingcited by 62
  22. 2025Step-3 is Large yet Affordable: Model-system Co-design for Cost-effective Decodingcited by 52
  23. 2024Small Language Model Meets with Reinforced Vision Vocabularycited by 43
  24. 2023Frozen Transformers in Language Models Are Effective Visual Encoder Layerscited by 39
  25. 2024OneChart: Purify the Chart Structural Extraction via One Auxiliary Tokencited by 39
  26. 2025POINTS-Reader: Distillation-Free Adaptation of Vision-Language Models for Document Conversioncited by 34
  27. 2026STEP3-VL-10B Technical Reportcited by 28
  28. 2024Slow Perception: Let's Perceive Geometric Figures Step-by-stepcited by 24
  29. 2025Context Cascade Compression: Exploring the Upper Limits of Text Compressioncited by 17

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 68.

68 papers.

Sort:

By area:

  1. 2026HunyuanOCR-1.5: Making Lightweight OCR VLMs Faster and Bettercited by 14
  2. 2026FireRed-OCR Technical Reportcited by 12
  3. 2026Multimodal OCR: Parse Anything from Documentscited by 12
  4. 2026OCRVerse: Towards Holistic OCR in End-to-End Vision-Language Modelscited by 10
  5. 2026Unlimited OCR Workscited by 8
  6. 2026Penguin-VL: Exploring the Efficiency Limits of VLM with LLM-based Vision Encoderscited by 7
  7. 2026UniVLR: Unifying Text and Vision in Visual Latent Reasoning for Multimodal LLMscited by 6
  8. 2026AgenticOCR: Parsing Only What You Need for Efficient Retrieval-Augmented Generationcited by 5
  9. 2026OvisOCR2 Technical Reportcited by 5
  10. 2026U-VLM: Hierarchical Vision Language Modeling for Report Generationcited by 4
  11. 2026Logics-Parsing-Omni Technical Reportcited by 4
  12. 2026MDPBench: A Benchmark for Multilingual Document Parsing in Real-World Scenarioscited by 4
38 more
  1. 2026RTPrune: Reading-Twice Inspired Token Pruning for Efficient DeepSeek-OCR Inferencecited by 4
  2. 2026Reading or Guessing? Visual Grounding Failures of Vision-Language Models for OCR in Ancient Greek Editionscited by 4
  3. 2026Infinity-Parser2 Technical Reportcited by 4
  4. 2026MonkeyOCRv2: A Visual-Text Foundation Model for Document AIcited by 4
  5. 2026SnapGuard: Lightweight Prompt Injection Detection for Screenshot-Based Web Agentscited by 3
  6. 2026MolEmb: Multimodal Large Language Models Can Be Strong Molecular Embedding Modelscited by 3
  7. 2026Data Science and Technology Towards AGI Part I: Tiered Data Managementcited by 2
  8. 2026Source or It Didn't Happen: A Multi-Agent Framework for Citation Hallucination Detectioncited by 2
  9. 2026P-MTP: Efficient Document Parsing via Multi-Token Prediction with Progressive Depth Scalingcited by 2
  10. 2026Advancing WordArt-Oriented Scene Text Recognition: Datasets and Methodscited by 2
  11. 2026StrucTab: A Structured Optimization Framework for Table Parsingcited by 2
  12. 2026Do VLMs Read or Rewrite? On Transcription Faithfulness in Vision-Language Modelscited by 2
  13. 2026Towards Universal Khmer Text Recognitioncited by 1
  14. 2026Towards Khmer Scene Document Layout Detectioncited by 1
  15. 2026LingDT-VL-OCR: Structure-Aware Document-Level Parsing with Fine-Grained Visual Referencecited by 1
  16. 2026MURE: Hierarchical Multi-Resolution Encoding via Vision-Language Models for Visual Document Retrievalcited by 1
  17. 2026Q-Mask: Query-driven Causal Masks for Text Anchoring in OCR-Oriented Vision-Language Modelscited by 1
  18. 2026Fine-tuning DeepSeek-OCR-2 for Molecular Structure Recognitioncited by 1
  19. 2026METATR: A Multilingual, Evolving Benchmark for Automatic Text Recognitioncited by 1
  20. 2026VTI-CoT: Visual-Textual Interleaved Chain of Thought for Video Reasoningcited by 1
  21. 2026Magnifying What Matters: Attention-Guided Adaptive Rendering for Visual Text Comprehensioncited by 1
  22. 2026LayoutLite: Token-Level Implicit Layout Analysis for Efficient Document OCRcited by 1
  23. 2026AndroidReality: How Far Are Mobile Agents from the Real World?cited by 1
  24. 2026FinixDoc: Rethinking Financial Document Parsing Beyond Saturated Benchmarkscited by 1
  25. 2026OCR-EDR: Rendering-Aware Diagnosis and Repair for Closed-Loop OCR Improvementcited by 1
  26. 2026All-in-One Multilingual Scene Text Recognition with Script-aware Mixture-of-Expertscited by 1
  27. 2026How to Utilize Complementary Vision-Text Information for 2D Structure Understandingcited by 0
  28. 2026Prefix-Adaptive Block Diffusion for Efficient Document Recognitioncited by 0
  29. 2026UniVL: Unified Vision-Language Embedding for Spatially Grounded Contextual Image Generationcited by 0
  30. 2026MPDocBench-Parse: Benchmarking Practical Multi-page Document Parsingcited by 0
  31. 2026GlobalDentBench: A Multinational Benchmark for Evaluating LLM Clinical Reasoning in Dentistry with Expert Calibrationcited by 0
  32. 2026Dr. DocBench: A Comprehensive Benchmark for Expert-Level and Difficult Document Parsingcited by 0
  33. 2026POTATR: A Lightweight Image-to-Graph Model for Page-Level Table Extractioncited by 0
  34. 2026LEDGER: A Long-Context Benchmark of Corporate Annual Reports for Grounded Financial Retrieval and Extractioncited by 0
  35. 2026PorTEXTO: A European Portuguese Benchmark for Visual Text Extractioncited by 0
  36. 2026Cross-Temporal Sinhala OCR: Page-Level Adaptation and Diachronic Analysiscited by 0
  37. 2026Why Struggle with Continuous Latents? Interpretable Discrete Latent Reasoning via Rendered Compressioncited by 0
  38. 2026HPD-Parsing: Hierarchical Parallel Document Parsingcited by 0

50 most cited of 68 All 68 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2024DeepSeek-VL: Towards Real-World Vision-Language Understanding
  2. 2026From Prompts to Tokens: Internalizing Causal Supervision in Vision-Language Model for Multi-Image Causal Reasoning
  3. 2025Reading Images Like Texts: Sequential Image Understanding in Vision-Language Models
  4. 2026Visual Merit or Linguistic Crutch? A Close Look at DeepSeek-OCR
  5. 2026Do Vision Language Models Need to Process Image Tokens?
  6. 2026Where To Look? : Causal Tracing of Vision Encoders in VLM
  7. 2026VisualScratchpad: Inference-time Visual Concepts Analysis in Vision Language Models
  8. 2025VScan: Rethinking Visual Token Reduction for Efficient Large Vision-Language Models