Jina AI launches jina-ocr-v1, a 570M-active-param visual document parser

JinaAI_ · x · 2026-09-18

Jina AI announced jina-ocr-v1, a new visual document parser with 3.4B total parameters and only 570M active parameters, with built-in speculative decoding. Throw PDFs, scans, tables, charts, or invoices at it and get clean markdown back.

It's available on Hugging Face today and via Jina Reader's x-respond-with parameter, targeting document understanding and RAG data-cleaning workflows.

Related event: Jina AI Releases Open-Source OCR Model jina-ocr-v1(3 posts)→

Original post →

More from Models

Models channel →