# Unlimited-OCR by BAIDU: Open-Weight Model Source: https://savrn.com/models/unlimited-ocr Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Runs On What it takes to serve Unlimited-OCR (3.3B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them. | Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits | | --- | --- | --- | --- | --- | --- | | 16-bit | 6.7 GB | 8.0 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 8-bit | 3.3 GB | 4.0 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 4-bit | 1.7 GB | 2.0 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the [SAVRN Index](https://savrn.com/ai-index/pricing/gpus), read Oct 24, 2026. [Unlimited-OCR on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/unlimited-ocr/gpus) ## SAVRN's Notes on Unlimited-OCR Unlimited-OCR is Baidu's document model, 3.3B parameters with a 32,768-token context, and one of the fastest-rising models in our catalog: about 2.3 million downloads a month and more than 4,200 likes within weeks of its June 2026 release. It needs about 8 GB at 16-bit, one mid-range GPU. The MIT license allows commercial use. For document parsing at scale, test it next to Chandra, PaddleOCR-VL and LightOnOCR on your own pages; the differences show up on tables and layouts. SAVRN Research, 2026-09-18 ## Model Card By BAIDU, published under mit, revision 8249d20b2822. ### Unlimited OCR Works #### Welcome the Era of One-shot Long-horizon Parsing. ### Release - [2026/07/21] Thanks to the[ms-swift community](https://github.com/modelscope/ms-swift) for their support, our model now supports training with [ms-swift](https://github.com/modelscope/ms-swift). - [2026/07/03] Thanks to the Baidu Cloud team for their support. Our model is now available on[Baidu Cloud](https://cloud.baidu.com/doc/OCR/s/fmr1p39gb). - [2026/06/28] Thanks to the[vLLM community](https://github.com/vllm-project/vllm) and [Tianyu Guo](https://github.com/gty111) for their support, our model now supports vLLM inference. - [2026/06/24] Thanks to[AK](https://x.com/_akhaliq) for creating a demo for us. It is now available at [Hugging Face Spaces](https://huggingface.co/spaces/baidu/Unlimited-OCR). - [2026/06/23] Our paper is now available on[arXiv](https://arxiv.org/abs/2606.23050). - [2026/06/23] Thanks to the[ModelScope community](https://github.com/modelscope) for their support. Our model is now available at [ModelScope](https://modelscope.cn/models/PaddlePaddle/Unlimited-OCR). - [2026/06/22] We present[Unlimited-OCR](https://github.com/baidu/Unlimited-OCR), aiming to push [Deepseek-OCR](https://github.com/deepseek-ai/DeepSeek-OCR) one step further. ### Inference #### Transformers Inference using Huggingface transformers on NVIDIA GPUs. Requirements tested on python 3.12.3 + CUDA12.9： ``` torch==2.10.0 torchvision==0.25.0 transformers==4.57.1 Pillow==12.1.1 matplotlib==3.10.8 einops==0.8.2 addict==2.4.0 easydict==1.13 pymupdf==1.27.2.2 psutil==7.2.2 ``` [Read the full model card (844 words)](https://savrn.com/models/unlimited-ocr/card) ## Configuration Architecture UnlimitedOCRForCausalLM Context length (tokens) 32,768 Layers 12 Hidden size 1,280 Feed-forward size 6,848 Attention heads 10 Key/value heads 10 Vocabulary size 129,280 Routed experts 64 Experts active per token 6 Sliding window (tokens) 128 Stored precision bfloat16 Model type unlimited_ocr ## Identity and Version Repository baidu/Unlimited-OCR Publisher BAIDU Task Image and text to text Modality Image and text Library transformers Parameters 3.3B parameters Languages ocr Revision 8249d20b2822a7d8a4c22c8d8c9069b5ac36bda3 First published 2026-06-19 Last updated 2026-10-14 ## Files and Weights 24 files, 13.5 GB in total. The weights are 2 files totalling 13.3 GB in safetensors. Weights2 files · 13.3 GB Configuration11 files · 464.0 KB Tokenizer2 files · 10.1 MB Documentation2 files · 12.2 KB Other6 files · 95.2 MB Repository1 file · 302 B Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | model-00001-of-000001.safetensors | Weights | 6.7 GB | 2bc48a7a1100 | | model.safetensors | Weights | 6.7 GB | 2bc48a7a1100 | | config.json | Configuration | 2.9 KB | — | | configuration_deepseek_v2.py | Configuration | 10.7 KB | — | | conversation.py | Configuration | 9.3 KB | — | | deepencoder.py | Configuration | 38.0 KB | — | | generation_config.json | Configuration | 202 B | — | | model.safetensors.index.json | Configuration | 257.6 KB | — | | modeling_deepseekv2.py | Configuration | 90.2 KB | — | | modeling_unlimitedocr.py | Configuration | 53.4 KB | — | | preprocessor_config.json | Configuration | 457 B | — | | processor_config.json | Configuration | 452 B | — | | special_tokens_map.json | Configuration | 801 B | — | | LICENSE | Documentation | 1.1 KB | — | | README.md | Documentation | 11.1 KB | — | | Unlimited-OCR.pdf | Other | 460.3 KB | d4cc0b2e98f5 | | assets/Unlimited-OCR.png | Other | 106.3 KB | 77063289aecf | | assets/baidu.png | Other | 11.1 KB | — | | assets/long-horizon-ocr.gif | Other | 82.2 MB | d4ba8964d33f | | chat_template.jinja | Other | 529 B | — | | wheel/sglang-0.0.0.dev11416+g92e8bb79e-py3-none-any.whl | Other | 12.5 MB | 2644a1f349c5 | | .gitattributes | Repository | 302 B | — | | tokenizer.json | Tokenizer | 10.0 MB | — | | tokenizer_config.json | Tokenizer | 165.9 KB | — | ## License and Download License mit Access Open weights, no gate Download size 13.3 GB [Download from BAIDU](https://huggingface.co/baidu/Unlimited-OCR) Released by BAIDU through its official repository on Hugging Face. [Read the license](https://opensource.org/license/mit). ## Built From - Described by arXiv:2606.23050 ## Evaluations Each result is shown as reported, with the conditions its reporter stated. None is a SAVRN measurement. A comparison lines two results up only when their configuration, unit and setup are all stated and identical. | Benchmark | Conditions | Result | Reported by | Revision | Date | | --- | --- | --- | --- | --- | --- | | llamaindex/ParseBench | Task chartMetric chartSetup Pipeline name: unlimitedocrComparison conditions not established | 1.34 | [ParseBench](https://huggingface.co/datasets/llamaindex/ParseBench) Reported by a third party | Evaluated revision not stated | 2026-06-24 | | llamaindex/ParseBench | Task layoutMetric layoutSetup Pipeline name: unlimitedocrComparison conditions not established | 71.52 | [ParseBench](https://huggingface.co/datasets/llamaindex/ParseBench) Reported by a third party | Evaluated revision not stated | 2026-06-24 | | llamaindex/ParseBench | Task meanMetric meanSetup Pipeline name: unlimitedocrComparison conditions not established | 46.17 | [ParseBench](https://huggingface.co/datasets/llamaindex/ParseBench) Reported by a third party | Evaluated revision not stated | 2026-06-24 | | llamaindex/ParseBench | Task tableMetric tableSetup Pipeline name: unlimitedocrComparison conditions not established | 70.21 | [ParseBench](https://huggingface.co/datasets/llamaindex/ParseBench) Reported by a third party | Evaluated revision not stated | 2026-06-24 | | llamaindex/ParseBench | Task text_contentMetric text_contentSetup Pipeline name: unlimitedocrComparison conditions not established | 86.81 | [ParseBench](https://huggingface.co/datasets/llamaindex/ParseBench) Reported by a third party | Evaluated revision not stated | 2026-06-24 | | llamaindex/ParseBench | Task text_formattingMetric text_formattingSetup Pipeline name: unlimitedocrComparison conditions not established | 0.97 | [ParseBench](https://huggingface.co/datasets/llamaindex/ParseBench) Reported by a third party | Evaluated revision not stated | 2026-06-24 | ## Memory Requirements | Precision | Weights in memory | | --- | --- | | As published | 13.3 GB | | 16-bit | 6.7 GB | | 8-bit | 3.3 GB | | 4-bit | 1.7 GB | Weights only, from the published parameter count; the key-value cache and runtime add to this. ## Built on This Model - Quantized from[Unlimited-OCR-AWQ](https://savrn.com/models/unlimited-ocr-awq) - Derived from[Unlimited-OCR-AWQ](https://savrn.com/models/unlimited-ocr-awq) - Quantized from[unlimited-ocr-mlx-coreml](https://savrn.com/models/unlimited-ocr-mlx-coreml) - Derived from[unlimited-ocr-mlx-coreml](https://savrn.com/models/unlimited-ocr-mlx-coreml) ## Compare Unlimited-OCR - [DeepSeek-OCR vs Unlimited-OCR](https://savrn.com/models/compare/deepseek-ocr-vs-unlimited-ocr) ## Questions About Unlimited-OCR ### How much GPU memory does Unlimited-OCR need? About 8 GB at 16-bit and 2 GB at 4-bit: the weights (3.3B parameters) plus a working margin. A long context needs more. ### What is the cheapest GPU to run Unlimited-OCR on? At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists. ### Can I use Unlimited-OCR commercially? Yes. Unlimited-OCR is released under MIT License. The MIT License is a short permissive license. It permits commercial use, modification and redistribution, provided the copyright notice and permission notice are included. ### What is Unlimited-OCR's context length? 32,768 tokens, from the maximum position embeddings in its published configuration. ## Similar Models Model · Image and text to text ### [DeepSeek-OCR](https://savrn.com/models/deepseek-ocr) [DeepSeek](https://savrn.com/model-publishers/deepseek-ai) Inference using Huggingface transformers on NVIDIA GPUs. Requirements tested on python 3.12.9 + CUDA11.8： Refer to GitHub for guidance on model inference acceleration and PDF processing, etc. [2025/10/23] DeepSeek-OCR is now officially supported in upstream vLLM. We would like to thank Vary, GOT-OCR2.0, MinerU, PaddleOCR, OneChart, Slow Perception for their valuable models and ideas. author={Wei, Haoran and Sun, Yaofeng and Li, Yukun}, year={2025} Open weights mit 3.3B parameters 8,192 tokens transformers [View model](https://savrn.com/models/deepseek-ocr) Model · Image and text to text ### [unlimited-ocr-max](https://savrn.com/models/unlimited-ocr-max) [Konstantin](https://savrn.com/model-publishers/kthierbach) baidu/Unlimited-OCR as one repository max serve opens directly on Apple Silicon: baidu's weights unchanged (model.safetensors, byte-identical to upstream at revision 07dea832e22aefee32ad281d4b80551282e1c168, sha256 2bc48a7a110061ea58fff65d3169367eebe3aee371ca6968dc2219c1b2855fc6), the tokenizer files as published, and the checkpoint's config.json with two keys removed (automap, modeltype) so MAX can load it without trustremotecode. The runtime is a MAX custom architecture with two Mojo custom ops, served from https://github.com/kthr/unlimited-ocr-max as an OpenAI-compatible endpoint on the Metal GPU or the CPU. The repository also carries model-int8.safetensors, this port's weight-only int8… Open weights mit 3.3B parameters 32,768 tokens [View model](https://savrn.com/models/unlimited-ocr-max) Model · Image and text to text ### [Unlimited-OCR-AWQ](https://savrn.com/models/unlimited-ocr-awq) [Sahil Chachra](https://savrn.com/model-publishers/sahilchachra) AWQ 4-bit (W4A16) quantization of baidu/Unlimited-OCR, a 3B vision-language OCR model that pushes DeepSeek-OCR one step further (one-shot, long-horizon document parsing). This repo quantizes the DeepSeek-V2 MoE text decoder with activation-aware scaling (AWQ) while keeping the vision tower in BF16, so it stays a drop-in transformers model. Unlimited-OCR uses the DeepSeek-OCR prompt vocabulary. The prompt must contain; prefix it with whenever you also want bounding boxes for what was read. - base — basesize=1024, imagesize=1024, cropmode=False. Good default for normal pages. - gundam — basesize=1024, imagesize=640, cropmode=True. Tiles the page; use for dense or large/high-resolution… Open weights mit 3.4B parameters 32,768 tokens transformers [View model](https://savrn.com/models/unlimited-ocr-awq) Model · Image and text to text ### [DeepSeek-OCR-2](https://savrn.com/models/deepseek-ocr-2) [DeepSeek](https://savrn.com/model-publishers/deepseek-ai) Inference using Huggingface transformers on NVIDIA GPUs. Requirements tested on python 3.12.9 + CUDA11.8： Refer to GitHub for guidance on model inference acceleration and PDF processing, etc. We would like to thank DeepSeek-OCR, Vary, GOT-OCR2.0, MinerU, PaddleOCR for their valuable models and ideas. We also appreciate the benchmark OmniDocBench. author={Wei, Haoran and Sun, Yaofeng and Li, Yukun}, year={2025} title={DeepSeek-OCR 2: Visual Causal Flow}, author={Wei, Haoran and Sun, Yaofeng and Li, Yukun}, year={2026} Open weights apache-2.0 3.4B parameters 8,192 tokens transformers [View model](https://savrn.com/models/deepseek-ocr-2) Model · Image and text to text ### [P110-ThaiOCR-4B-NVFP4](https://savrn.com/models/p110-thaiocr-4b-nvfp4) [Porsuk Rungarunsrisuk](https://savrn.com/model-publishers/porsukr) P110-ThaiOCR-4B is a 4B vision-language model for Thai documents: printed and handwritten text, tables, forms, government PDFs and scans. It reads pages into plain text or structured output and can answer questions about a document (key information, classification, charts), without per-template rules. NVFP4 for vLLM on NVIDIA Blackwell: MLP in NVFP4, attention in FP8, compressed-tensors format. Recommended prompt for plain transcription (Thai): Tested with vLLM 0.29.0 and transformers 5.17.0. All open models were run on the same hardware with the same scoring code (October 2026). Qwen3.5-4B is the base model P110 was fine-tuned from. Models with two runs are averaged. Bold marks the best… Open weights apache-2.0 3.4B parameters 262,144 tokens transformers [View model](https://savrn.com/models/p110-thaiocr-4b-nvfp4) Model · Image and text to text ### [Qwen3.5-4B-W4A16-AutoRound-NVFP4](https://savrn.com/models/qwen3-5-4b-w4a16-autoround-nvfp4) [Vishva R](https://savrn.com/model-publishers/vishva007) This is an NVFP4 (W4A16: 4-bit weight, 16-bit activation) quantized version of Qwen/Qwen3.5-4B, produced using AutoRound — Intel's sign gradient descent based quantization method designed for production-grade accuracy retention. Exported via compressed-tensors (llmcompressor format) with MTP layers preserved. - NVFP4 Quantization — Weights are quantized to FP4 micro-scaled formats using group size 16; activations remain in 16-bit for stable numerical fidelity. - Vision Tower (visual / quantnontextmodule): Kept in BF16 to preserve visual reasoning and OCR precision. - Special Modules (ignorelayers & layerconfig): linearattn, mtp, and mtp.fc are strictly retained in native bfloat16 to ensure… Open weights apache-2.0 3.4B parameters 262,144 tokens [View model](https://savrn.com/models/qwen3-5-4b-w4a16-autoround-nvfp4) ## BAIDU [All models and datasets](https://savrn.com/model-publishers/baidu) ## Versions - [8249d20b2822](https://savrn.com/models/unlimited-ocr/versions/8249d20b2822) · current 2026-10-24 - [07dea832e22a](https://savrn.com/models/unlimited-ocr/versions/07dea832e22a) 2026-10-14 ## Explore More - [All image and text to text models](https://savrn.com/models/tasks/image-and-text-to-text) - [All models under mit](https://savrn.com/models/licenses/mit) - [Model comparisons](https://savrn.com/models/comparisons) - [The model directory](https://savrn.com/models) - [Open model prices by host](https://savrn.com/ai-index/pricing/open-models) ## Source - Repository metadata, read 2026-10-15. - [Hugging Face record](https://huggingface.co/baidu/Unlimited-OCR) - [How the hub is built](https://savrn.com/model-hub/methodology)