selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

Alternatives to Unlimited-OCR-2

120 open-weight models that do the same job as Unlimited-OCR-2 (3.3B OCR), ranked by how close they are in task and size, then by popularity and how recent they are. Licence and VRAM are shown for each.

the original
baidu logo

Unlimited-OCR-2

baidu·OCR

≈4.5 GB

Sparse Reference Sliding Window Attention (SR-SWA): Compresses text-visual attention interactions from ultra-long sequences down to a fixed-size set of working references (e.g., 512 tokens).

3.3BA1.1B128K ctx MIT
612.1KOct 2026gguf
--vram
any
--gpus
--quant
  1. opendatalab logo

    MinerU2.5-Pro-2604-1.2B

    opendatalab·OCR

    ≈3.1 GB

    MinerU2.5-Pro is our latest document parsing model (PDF-to-Markdown) that establishes a new industry standard. By focusing entirely on data engineering without altering the original 1.2B-parameter architecture, it delivers exceptional results across the board:

    1.2B8K ctx Apache 2.0
    27.8KApr 2026gguf
    • Same category
    • Smaller (1.2B vs 3.3B)
    • Commercial use OK
    vs
  2. PaddlePaddle logo

    PaddleOCR-VL-1.6

    PaddlePaddle·OCR

    ≈2.6 GB

    PaddleOCR-VL-1.6: Expanding the Frontier of Document Parsing with Under-Optimized Region Refinement and Progressive Post-Training

    959M128K ctx Apache 2.0
    96.9KMay 2026gguf
    • Same category
    • Smaller (959M vs 3.3B)
    • Commercial use OK
    vs
  3. opendatalab logo

    MinerU2.5-2509-1.2B

    opendatalab·OCR

    ≈3.1 GB

    MinerU2.5 is a 1.2B-parameter vision-language model for document parsing that achieves state-of-the-art accuracy with high computational efficiency.

    1.2B16K ctx AGPL 3.0
    77.7KSept 2025gguf
    • Same category
    • Smaller (1.2B vs 3.3B)
    • Commercial use OK
    vs
  4. datalab-to logo

    surya-ocr-2

    datalab-to·OCR

    ≈2.0 GB

    It works on a range of documents (see usage and benchmarks).

    686M256K ctx OpenRAIL
    615.2KMay 2026gguf
    • Same category
    • Smaller (686M vs 3.3B)
    vs
  5. datalab-to logo

    surya-ocr-2.5

    datalab-to·OCR

    ≈2.0 GB

    We also ship smaller models for line-level text detection and ocr error detection. It works on a range of documents (see usage and benchmarks).

    689M256K ctx OpenRAIL
    136.1KOct 2026gguf
    • Same category
    • Smaller (689M vs 3.3B)
    vs
  6. lightonai logo

    LightOnOCR-3-1B

    lightonai·OCR

    ≈2.8 GB

    Same architecture as LightOnOCR-2-1B, new capabilities. LightOnOCR-3-1B keeps the architecture of LightOn's previous LightOnOCR-2 models, so switching requires no change, and adds the new grounding, image description and chart extraction features.

    1B16K ctx Apache 2.0
    4.4KOct 2026gguf
    • Same category
    • Smaller (1B vs 3.3B)
    • Commercial use OK
    vs
  7. allenai logo≈10 GB

    This is a preview release of the olmOCR model that's fine tuned from Qwen2-VL-7B-Instruct using the olmOCR-mix-0225 dataset.

    8.3B32K ctx Apache 2.0
    33.8KJan 2025gguf
    • Same category
    • Larger (8.3B vs 3.3B)
    • Commercial use OK
    vs
  8. lightonai logo

    LightOnOCR-1B-1025

    lightonai·OCR

    ≈3.1 GB

    Full BF16 version of the model. We recommend this variant for inference and further fine-tuning.

    1.2B8K ctx Apache 2.0
    14.4KOct 2025gguf
    • Same category
    • Smaller (1.2B vs 3.3B)
    • Commercial use OK
    vs
  9. lightonai logo

    LightOnOCR-3-0.8B

    lightonai·OCR

    ≈2.4 GB

    Smallest model. LightOnOCR-3-0.8B is the lightest model of LightOn's LightOnOCR-3 family, with the same layout and visual understanding features.

    853M256K ctx Apache 2.0
    15KOct 2026gguf
    • Same category
    • Smaller (853M vs 3.3B)
    • Commercial use OK
    vs
  10. ATH-MaaS logo

    OvisOCR2

    ATH-MaaS·OCR

    ≈2.4 GB

    We are pleased to announce the release of OvisOCR2, a compact 0.8B end-to-end model for page-level document parsing. Given a document page image, OvisOCR2 generates a Markdown representation in natural reading order, covering text, formulas, tables, and visual regions.

    853M256K ctx Apache 2.0
    24.4KJul 2026gguf
    • Same category
    • Smaller (853M vs 3.3B)
    • Commercial use OK
    vs
  11. PaddlePaddle logo

    PaddleOCR-VL-1.5

    PaddlePaddle·OCR

    ≈2.6 GB

    PaddleOCR-VL-1.5: Towards a Multi-Task 0.9B VLM for Robust In-the-Wild Document Parsing

    959M128K ctx Apache 2.0
    16.3KJan 2026gguf
    • Same category
    • Smaller (959M vs 3.3B)
    • Commercial use OK
    vs
  12. datalab-to logo

    chandra

    datalab-to·OCR

    ≈11 GB

    Chandra is an OCR model that outputs markdown, HTML, and JSON. It is highly accurate at extracting text from images and PDFs, while preserving layout information.

    8.8B256K ctx OpenRAIL
    1.9KOct 2025gguf
    • Same category
    • Larger (8.8B vs 3.3B)
    vs
  13. nanonets logo≈4.2 GB

    Nanonets-OCR2 by Nanonets is a family of powerful, state-of-the-art image-to-markdown OCR models that go far beyond traditional text extraction.

    1.6B32K ctx Apache 2.0
    187Oct 2025gguf
    • Same category
    • Smaller (1.6B vs 3.3B)
    • Commercial use OK
    vs
  14. inclusionAI logo

    ArmorOCR

    inclusionAI·OCR

    ≈11 GB

    ArmorOCR is a two-stage framework for grounded adversarial OCR perception built on Qwen3-VL-8B-Instruct. It enables single-pass inference on the original image, without any inference-time visual transformations or tool assistance.

    8.8B256K ctx Apache 2.0
    154Aug 2026gguf
    • Same category
    • Larger (8.8B vs 3.3B)
    • Commercial use OK
    vs
  15. PaddlePaddle logo

    PaddleOCR-VL

    PaddlePaddle·OCR

    ≈2.6 GB

    PaddleOCR-VL: Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language Model

    959M128K ctx Apache 2.0
    18.8KOct 2025gguf
    • Same category
    • Smaller (959M vs 3.3B)
    • Commercial use OK
    vs
  16. lightonai logo

    LightOnOCR-2-1B-base

    lightonai·OCR

    ≈2.8 GB

    Base model for fine-tuning. This is the pre-RLVR checkpoint with strong OCR capabilities, ideal as a starting point for domain adaptation and custom fine-tuning.

    1B16K ctx Apache 2.0
    3.6KJan 2026gguf
    • Same category
    • Smaller (1B vs 3.3B)
    • Commercial use OK
    vs
  17. PaddlePaddle logo

    PP-OCRv6_medium_det

    PaddlePaddle·OCR

    ? GB

    PP-OCRv6 is a lightweight OCR system that combines architectural innovation with data-centric optimization. It redesigns the backbone, detection neck, and recognition neck around a unified MetaFormer-style building block with structural reparameterization.

    ?BApache 2.0
    48.5KJun 2026
    • Same category
    • Commercial use OK
    vs
  18. allenai logo

    olmOCR-7B-0725

    allenai·OCR

    ≈10 GB

    This is a release of the olmOCR model that's fine tuned from Qwen2.5-VL-7B-Instruct using the olmOCR-mix-0225 dataset.

    8.3B125K ctx Apache 2.0
    779Jul 2025gguf
    • Same category
    • Larger (8.3B vs 3.3B)
    • Commercial use OK
    vs
  19. allenai logo

    olmOCR-7B-0825

    allenai·OCR

    ≈10 GB

    This is a release of the olmOCR model that's fine tuned from Qwen2.5-VL-7B-Instruct using the olmOCR-mix-0225 dataset.

    8.3B125K ctx Apache 2.0
    684Aug 2025gguf
    • Same category
    • Larger (8.3B vs 3.3B)
    • Commercial use OK
    vs
  20. lightonai logo

    LightOnOCR-2-1B-bbox

    lightonai·OCR

    ≈2.8 GB

    Best model with image localization. This model predicts both text and precise bounding boxes for embedded images, refined with RLVR using IoU-based rewards for maximum localization accuracy.

    1B16K ctx Apache 2.0
    3KJan 2026gguf
    • Same category
    • Smaller (1B vs 3.3B)
    • Commercial use OK
    vs
  21. reducto logo

    RolmOCR

    reducto·OCR

    ≈10 GB

    Earlier this year, the Allen Institute for AI released olmOCR, an open-source tool that performs document OCR using the Qwen2-VL-7B vision language model (VLM).

    8.3B125K ctx Apache 2.0
    1.4KApr 2025gguf
    • Same category
    • Larger (8.3B vs 3.3B)
    • Commercial use OK
    vs
  22. PaddlePaddle logo

    PP-OCRv5_server_det

    PaddlePaddle·OCR

    ? GB

    PP-OCRv5_server_det is one of the PP-OCRv5_det series, the latest generation of text detection models developed by the PaddleOCR team.

    ?BApache 2.0
    452.6KJun 2025
    • Same category
    • Commercial use OK
    vs
  23. lightonai logo≈2.8 GB

    Merged bbox variant (advanced). This model combines OCR-improving RLVR signals with bounding-box-focused RLVR updates via joint merging, preserving OCR quality while providing image localization.

    1B16K ctx Apache 2.0
    1.4KJan 2026gguf
    • Same category
    • Smaller (1B vs 3.3B)
    • Commercial use OK
    vs
  24. PaddlePaddle logo

    en_PP-OCRv5_mobile_rec

    PaddlePaddle·OCR

    ? GB

    en_PP-OCRv5_mobile_rec is one of the PP-OCRv5_rec that are the latest generation text line recognition models developed by PaddleOCR team. It aims to efficiently and accurately support the recognition of English. The key accuracy metrics are as follow:

    ?BApache 2.0
    187.2KAug 2025
    • Same category
    • Commercial use OK
    vs