Unlimited OCR 2

Use it to extract and parse text from images and long documents, including PDFs and multi-page files.

Availability

Model available on Hugging Face, ModelScope, and Baidu Cloud

Pricing model

Not established

Deployment

Supports Transformers and SGLang inference

Privacy

Not established

Licence

MIT

• SR-SWA compresses text-visual attention interactions to a fixed-size set of working references, such as 512 tokens. ^[1]

  * 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).

• The model supports multi-page inference, with all pages of a document parsed in one request. ^[1]

  All pages of a document are parsed in one request, so the model can use cross-page context.

• The release describes unlimited-horizon parsing within a finite context window. ^[1]

  * From Long- to Unlimited-Horizon Parsing: Powered by SR-SWA, Unlimited-OCR 2 achieves true, unbounded document parsing within a finite context window—effectively eliminating quadratic attention scaling and memory bottlenecks as documents grow.

• The model is available on Hugging Face, ModelScope, and Baidu Cloud. ^[1]

  It is now available at [Hugging Face Spaces](https://huggingface.co/spaces/baidu/Unlimited-OCR).

• The model supports Transformers and SGLang inference. ^[1]

  - [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.

• The model is released under the MIT licence. ^[1]

• New models bring live speech translation and voice features Daily brief 2026-10-28