mlsanityData to 14 Oct 2026

MinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing

arXiv 2509.22186 ↗First posted 26 Sep 2025cs.CV cs.CLcited by 116

On the map: Multimodal LLMs / vision-language models → Document understanding and multimodal RAG

Abstract (arXiv)

We introduce MinerU2.5, a 1.2B-parameter document parsing vision-language model that achieves state-of-the-art recognition accuracy while maintaining exceptional computational efficiency. Our approach employs a coarse-to-fine, two-stage parsing strategy that decouples global layout analysis from local content recognition. In the first stage, the model performs efficient layout analysis on downsampled images to identify structural elements, circumventing the computational overhead of processing high-resolution inputs. In the second stage, guided by the global layout, it performs targeted content recognition on native-resolution crops extracted from the original image, preserving fine-grained details in dense text, complex formulas, and tables. To support this strategy, we developed a comprehensive data engine that generates diverse, large-scale training corpora for both pretraining and fine-tuning. Ultimately, MinerU2.5 demonstrates strong document parsing ability, achieving state-of-the-art performance on multiple benchmarks, surpassing both general-purpose and domain-specific models across various recognition tasks, while maintaining significantly lower computational overhead.

Key facts. Cited by 116 papers in the map (data through 14 Oct 2026). Placed in Document understanding and multimodal RAG, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 33 of its 49 references that are in the map. First posted on arXiv 26 Sep 2025.

Builds on

33 of 49 references.

Closest ancestors · 6

Same direction, newest first

  1. 2025POINTS-Reader: Distillation-Free Adaptation of Vision-Language Models for Document Conversion1 Sep
  2. 2025SmolDocling: An ultra-compact vision-language model for end-to-end multi-modal document conversion14 Mar
  3. 2025olmOCR: Unlocking Trillions of Tokens in PDFs with Vision Language Models25 Feb
  4. 2024OCRBench v2: An Improved Benchmark for Evaluating Large Multimodal Models on Visual Text Localization and Reasoning31 Dec
  5. 2024HRVDA: High-Resolution Visual Document Assistant10 Apr
  6. 2024TextMonkey: An OCR-Free Large Multimodal Model for Understanding Document7 Mar

Other references · 27

  1. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,944
  2. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,703
  3. 2024Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolutioncited by 4,234
  4. 2025InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiencycited by 1,510
  5. 2016Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Networkcited by 1,204
  6. 2024YOLOv10: Real-Time End-to-End Object Detectioncited by 515
  7. 2022LayoutLMv3: Pre-training for Document AI with Unified Text and Image Maskingcited by 485
  8. 2024Retrieval-Augmented Generation for AI-Generated Content: A Surveycited by 354
  9. 2024MinerU: An Open-Source Solution for Precise Document Content Extractioncited by 267
  10. 2023Nougat: Neural Optical Understanding for Academic Documentscited by 245
  11. 2019Image-based table recognition: data, model, and evaluationcited by 237
  12. 2023Patch n' Pack: NaViT, a Vision Transformer for any Aspect Ratio and Resolutioncited by 229
15 more
  1. 2022DocLayNet: A Large Human-Annotated Dataset for Document-Layout Analysiscited by 148
  2. 2020Global Table Extractor (GTE): A Framework for Joint Table Identification and Cell Structure Recognition Using Visual Contextcited by 147
  3. 2022Unifying Vision, Text, and Layout for Universal Document Processingcited by 140
  4. 2025MonkeyOCR: Document Parsing with a Structure-Recognition-Relation Triplet Paradigmcited by 87
  5. 2024DocLayout-YOLO: Enhancing Document Layout Analysis through Diverse Synthetic Data and Global-to-Local Adaptive Perceptioncited by 82
  6. 2021LayoutReader: Pre-training of Text and Layout for Reading Order Detectioncited by 80
  7. 2024OmniParser: A Unified Framework for Text Spotting, Key Information Extraction and Table Recognitioncited by 74
  8. 2023Vision Grid Transformer for Document Layout Analysiscited by 47
  9. 2024Image Over Text: Transforming Formula Recognition Evaluation with Character Detection Matchingcited by 41
  10. 2022VRDU: A Benchmark for Visually-rich Document Understandingcited by 35
  11. 2023Optimized Table Tokenization for Table Structure Recognitioncited by 34
  12. 2022Wukong-Reader: Multi-modal Pre-training for Fine-grained Visual Document Understandingcited by 20
  13. 2023DocTr: Document Transformer for Structured Information Extraction in Documentscited by 18
  14. 2024Revolutionizing Retrieval-Augmented Generation with Enhanced PDF Structure Recognitioncited by 16
  15. 2025PP-FormulaNet: Bridging Accuracy and Efficiency in Advanced Formula Recognitioncited by 7

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 17; 2026: 99.

116 papers.

Sort:

By area:

  1. 2025PaddleOCR-VL: Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language Modelcited by 89
  2. 2026PaddleOCR-VL-1.5: Towards a Multi-Task 0.9B VLM for Robust In-the-Wild Document Parsingcited by 57
  3. 2025olmOCR 2: Unit Test Rewards for Document OCRcited by 47
  4. 2026MMFineReason: Closing the Multimodal Reasoning Gap via Open Data-Centric Methodscited by 30
  5. 2026Intern-S1-Pro: Scientific Multimodal Foundation Model at Trillion Scalecited by 29
  6. 2026MinerU2.5-Pro: Pushing the Limits of Data-Centric Document Parsing at Scalecited by 27
  7. 2026PaperBanana: Automating Academic Illustration for AI Scientistscited by 26
  8. 2026GLM-OCR Technical Reportcited by 26
  9. 2025DataFlow: An LLM-Driven Framework for Unified Data Preparation and Workflow Automation in the Era of Data-Centric AIcited by 23
  10. 2026Dolphin-v2: Universal Document Parsing via Scalable Anchor Promptingcited by 22
  11. 2026PaddleOCR-VL-1.6: Expanding the Frontier of Document Parsing with Under-Optimized Region Refinement and Progressive Post-Trainingcited by 21
  12. 2025HunyuanOCR Technical Reportcited by 20
38 more
  1. 2025Resolving Evidence Sparsity: Agentic Context Engineering for Long-Document Understandingcited by 18
  2. 2025Nex-N1: Agentic Models Trained via a Unified Ecosystem for Large-Scale Environment Constructioncited by 17
  3. 2025DocLens : A Tool-Augmented Multi-Agent Framework for Long Visual Document Understandingcited by 15
  4. 2026Multimodal OCR: Parse Anything from Documentscited by 14
  5. 2026HunyuanOCR-1.5: Making Lightweight OCR VLMs Faster and Bettercited by 14
  6. 2025MonkeyOCR v1.5 Technical Report: Unlocking Robust Document Parsing for Complex Patternscited by 13
  7. 2025UniRec-0.1B: Unified Text and Formula Recognition with 0.1B Parameterscited by 13
  8. 2026FireRed-OCR Technical Reportcited by 13
  9. 2026OCRVerse: Towards Holistic OCR in End-to-End Vision-Language Modelscited by 12
  10. 2026Real5-OmniDocBench: A Full-Scale Physical Reconstruction Benchmark for Robust Document Parsing in the Wildcited by 12
  11. 2026DECKBench: Benchmarking Multi-Agent Frameworks for Academic Slide Generation and Editingcited by 11
  12. 2026ChartVerse: Scaling Chart Reasoning via Reliable Programmatic Synthesis from Scratchcited by 10
  13. 2026Youtu-Parsing: Perception, Structuring and Recognition via High-Parallelism Decodingcited by 10
  14. 2025DocPTBench: Benchmarking End-to-End Photographed Document Parsing and Translationcited by 9
  15. 2025MindGPT-4ov: An Enhanced MLLM via a Multi-Stage Post-Training Paradigmcited by 9
  16. 2025Reading or Reasoning? Format Decoupled Reinforcement Learning for Document OCRcited by 9
  17. 2026MinerU-Diffusion: Rethinking Document OCR as Inverse Rendering via Diffusion Decodingcited by 8
  18. 2026Visual Late Chunking: An Empirical Study of Contextual Chunking for Efficient Visual Document Retrievalcited by 8
  19. 2026Towards Real-World Document Parsing via Realistic Scene Synthesis and Document-Aware Trainingcited by 7
  20. 2025SO-Bench: A Structural Output Evaluation of Multimodal LLMscited by 6
  21. 2025DOCR-Inspector: Fine-Grained and Automated Evaluation of Document Parsing with VLMcited by 6
  22. 2026HSD: Training-Free Acceleration for Document Parsing Vision-Language Models with Hierarchical Speculative Decodingcited by 6
  23. 2026SVCBench: A Streaming Video Counting Benchmark for Spatial-Temporal State Maintenancecited by 6
  24. 2026Beyond NL2Code: A Structured Survey of Multimodal Code Intelligencecited by 6
  25. 2026AgenticOCR: Parsing Only What You Need for Efficient Retrieval-Augmented Generationcited by 5
  26. 2026Logics-Parsing-Omni Technical Reportcited by 5
  27. 2026MDPBench: A Benchmark for Multilingual Document Parsing in Real-World Scenarioscited by 5
  28. 2026CiteVQA: Benchmarking Evidence Attribution for Trustworthy Document Intelligencecited by 5
  29. 2025TRivia: Self-supervised Fine-tuning of Vision-Language Models for Table Recognitioncited by 4
  30. 2026Beyond the Grid: Layout-Informed Multi-Vector Retrieval with Parsed Visual Document Representationscited by 4
  31. 2026MMTIT-Bench: A Multilingual and Multi-Scenario Benchmark with Cognition-Perception-Reasoning Guided Text-Image Machine Translationcited by 4
  32. 2026PRISM: Probing Reasoning, Instruction, and Source Memory in LLM Hallucinationscited by 4
  33. 2026How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settingscited by 4
  34. 2026MechVQA: Benchmarking and Enhancing Multimodal LLMs on Comprehensive Mechanical Drawing Understandingcited by 4
  35. 2026Infinity-Parser2 Technical Reportcited by 4
  36. 2026MonkeyOCRv2: A Visual-Text Foundation Model for Document AIcited by 4
  37. 2026NERFIFY: A Multi-Agent Framework for Turning NeRF Papers into Codecited by 3
  38. 2026Visual-ERM: Reward Modeling for Visual Equivalencecited by 3

50 most cited of 116 All 116 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026MinerU-Popo: Universal Post-Processing Model for Structured Document Parsing
  2. 2025dots.ocr: Multilingual Document Layout Parsing in a Single Vision-Language Model
  3. 2026Structured Layout Priors for Robust Out-of-Distribution Visual Document Understanding
  4. 2026Dr. DocBench: A Comprehensive Benchmark for Expert-Level and Difficult Document Parsing
  5. 2024Document Parsing Unveiled: Techniques, Challenges, and Prospects for Structured Information Extraction
  6. 2026P-MTP: Efficient Document Parsing via Multi-Token Prediction with Progressive Depth Scaling
  7. 2025Index-Preserving Lightweight Token Pruning for Efficient Document Understanding in Vision-Language Models
  8. 2026Read, Parse, Describe: Unified Document Parsing with Visual Element Description Generation