---
type: "post"
id: "baidu-open-sources-unlimited-ocr-2-a-3b-b8360f"
slug: "baidu-open-sources-unlimited-ocr-2-a-3b-b8360f"
title: "Baidu open-sources Unlimited OCR 2, a 3B MIT-licensed OCR model that parses documents of any length"
description: "Baidu released Unlimited OCR 2 on Hugging Face on October 27, 2026."
retrieval_nugget: "Baidu released Unlimited OCR 2 on Hugging Face on October 27, 2026."
published_at: "2026-10-27"
updated_at: "2026-10-27"
record_date: "2026-10-27"
date_kind: "published_at"
topics: ["open-models","multimodal"]
entities: ["Baidu","Unlimited OCR 2","Hugging Face","Transformers","SGLang","vLLM","MPDocBench","MarathonDocBench"]
related_patterns: []
novelty_label: "new_fact"
what_changed: "Baidu has released Unlimited OCR 2, a 3B-parameter MIT-licensed model that parses documents of any length."
prior_refs: ["memory:d5a0da2e-9dab-87db-ab2a-11b5658e776c","memory:b048da70-f832-8675-aee0-7733178725be"]
source_url: "https://huggingface.co/baidu/Unlimited-OCR-2"
source_urls: ["https://huggingface.co/baidu/Unlimited-OCR-2"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:baidu-open-sources-unlimited-ocr-2-a-3b-b8360f"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/baidu-open-sources-unlimited-ocr-2-a-3b-b8360f/","markdown":"https://newruntime.com/posts/baidu-open-sources-unlimited-ocr-2-a-3b-b8360f.md","json":"https://newruntime.com/posts/baidu-open-sources-unlimited-ocr-2-a-3b-b8360f.json"}
---

# Baidu open-sources Unlimited OCR 2, a 3B MIT-licensed OCR model that parses documents of any length

## Retrieval answer

Baidu released Unlimited OCR 2 on Hugging Face on October 27, 2026.

Baidu released Unlimited OCR 2 on Hugging Face on October 27, 2026. The public model repository is MIT-licensed, ungated, and identifies the model as an image-to-text system using Transformers custom code. Its safetensors metadata lists 3.34 billion BF16 parameters and roughly 6.69 GB of model files.

The release introduces Sparse Reference Sliding Window Attention (SR-SWA). Baidu says it compresses text-visual attention for ultra-long sequences into a fixed-size set of working references, with 512 tokens given as an example. The model card says this keeps per-step decoding latency and memory footprint nearly constant as document length grows, and supports multi-page input in one request so it can use cross-page context.

The repository documents local use through Transformers and recommends SGLang for serving. It also includes vLLM-oriented instructions. Baidu claims state-of-the-art results on MPDocBench and MarathonDocBench, including hundred-page document benchmarks.

The source does not independently verify those benchmark claims, establish accuracy on particular scans or languages, or provide a comparison of operating cost and hardware requirements across document sizes.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Open Models](https://newruntime.com/topics/open-models.md): Explore the open-models topic hub.
- topic hub: [Multimodal](https://newruntime.com/topics/multimodal.md): Explore the multimodal topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)