mlsanityData to 14 Oct 2026

LightOnOCR: A 1B End-to-End Multilingual Vision-Language Model for State-of-the-Art OCR

arXiv 2601.14251 ↗First posted 20 Jan 2026cs.CVcited by 23

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

Abstract (arXiv)

We present LightOnOCR-2-1B, a 1B-parameter end-to-end multilingual vision--language model that converts document images (e.g., PDFs) into clean, naturally ordered text without brittle OCR pipelines. Trained on a large-scale, high-quality distillation mix with strong coverage of scans, French documents, and scientific PDFs, LightOnOCR-2 achieves state-of-the-art results on OlmOCR-Bench while being 9× smaller and substantially faster than prior best-performing models. We further extend the output format to predict normalized bounding boxes for embedded images, introducing localization during pretraining via a resume strategy and refining it with RLVR using IoU-based rewards. Finally, we improve robustness with checkpoint averaging and task-arithmetic merging. We release model checkpoints under Apache 2.0, and publicly release the dataset and LightOnOCR-bbox-bench evaluation under their respective licenses.

Key facts. Cited by 23 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 15 of its 27 references that are in the map. First posted on arXiv 20 Jan 2026.

Builds on

15 of 27 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025dots.ocr: Multilingual Document Layout Parsing in a Single Vision-Language Model2 Dec
  2. 2025PaddleOCR-VL: Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language Model16 Oct
  3. 2025olmOCR: Unlocking Trillions of Tokens in PDFs with Vision Language Models25 Feb

Other references · 12

  1. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  2. 2024Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolutioncited by 4,234
  3. 2023FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioningcited by 2,598
  4. 2022Editing Models with Task Arithmeticcited by 1,120
  5. 2024Tulu 3: Pushing Frontiers in Open Language Model Post-Trainingcited by 1,053
  6. 2015An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognitioncited by 520
  7. 2021TrOCR: Transformer-based Optical Character Recognition with Pre-trained Modelscited by 279
  8. 2024MinerU: An Open-Source Solution for Precise Document Content Extractioncited by 267
  9. 2023Nougat: Neural Optical Understanding for Academic Documentscited by 245
  10. 2020PP-OCR: A Practical Ultra Lightweight OCR Systemcited by 151
  11. 2025olmOCR 2: Unit Test Rewards for Document OCRcited by 47
  12. 2025'Eclair -- Extracting Content and Layout with Integrated Reading Order for Documentscited by 9

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 23.

23 papers.

Sort:

By area:

  1. 2026Structure-Aware Text Recognition for Ancient Greek Critical Editionscited by 5
  2. 2026Reading or Guessing? Visual Grounding Failures of Vision-Language Models for OCR in Ancient Greek Editionscited by 5
  3. 2026RTPrune: Reading-Twice Inspired Token Pruning for Efficient DeepSeek-OCR Inferencecited by 4
  4. 2026Infinity-Parser2 Technical Reportcited by 4
  5. 2026Beyond String Matching: Semantic Evaluation of PDF Table Extractioncited by 2
  6. 2026GlotOCR Bench: OCR Models Still Struggle Beyond a Handful of Unicode Scriptscited by 2
  7. 2026P-MTP: Efficient Document Parsing via Multi-Token Prediction with Progressive Depth Scalingcited by 2
  8. 2026StrucTab: A Structured Optimization Framework for Table Parsingcited by 2
  9. 2026Benchmarking Vision-Language Models for French PDF-to-Markdown Conversioncited by 1
  10. 2026METATR: A Multilingual, Evolving Benchmark for Automatic Text Recognitioncited by 1
  11. 2026End-to-End Text Line Detection and Orderingcited by 0
  12. 2026Freeing the Law with LOCUS: A Local Ordinance Corpus for the United Statescited by 0
11 more
  1. 2026Cross-Temporal Sinhala OCR: Page-Level Adaptation and Diachronic Analysiscited by 0
  2. 2026ClinOCR-Bench: A Comprehensive Clinical Scanned Document Dataset for Optical Character Recognition Model Evaluationcited by 0
  3. 2026HPD-Parsing: Hierarchical Parallel Document Parsingcited by 0
  4. 2026PUMA: A Polish Benchmark for Culturally Grounded Multimodal Understandingcited by 0
  5. 2026OCR-Based Field Extraction for Archaeological Pottery Metadata: The CENTURIA Datasetcited by 0
  6. 2026Jina-OCR-v1: Efficient Document Parsing with Speculative Decoding and Dense Verifiable Rewardscited by 0
  7. 2026ExpertHTR: Unified Handwritten Text Recognition with Multi-Task Learning and Sparse Mixture-of-Expertscited by 0
  8. 2026A Free Lunch? Adapting PP-OCRv6 for Historical Text Recognitioncited by 0
  9. 2026The Spoken Wikipedia Presentation Corpuscited by 0
  10. 2026Xiaomi-OCR-0 Technical Reportcited by 0
  11. 2026The Last Mile of Coding Agents for Long-Document Tasks: An Empirical Study on Clinical Trial Protocol Digitizationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026PP-OCRv5: A Specialized 5M-Parameter Model Rivaling Billion-Parameter Vision-Language Models on OCR Tasks
  2. 2026Multilingual OCR-Aware Fine-Tuning and Prompt-Guided Chain-of-Thought Reasoning for Multimodal Large Language Models
  3. 2026Do VLMs Read or Rewrite? On Transcription Faithfulness in Vision-Language Models
  4. 2024OCRBench v2: An Improved Benchmark for Evaluating Large Multimodal Models on Visual Text Localization and Reasoning
  5. 2025DianJin-OCR-R1: Enhancing OCR Capabilities via a Reasoning-and-Tool Interleaved Vision-Language Model
  6. 2025E-ARMOR: Edge case Assessment and Review of Multilingual Optical Character Recognition
  7. 2026What Can Low Resource Languages Learn From Each Other?
  8. 2024TAP-VL: Text Layout-Aware Pre-training for Enriched Vision-Language Models