---
name: Deepseek Deepseek-Ocr-2
slug: ocr-2
type: Multimodal
providers:
  - novita-ai
description: OCR model for extracting structured text from documents and screenshots
tags:
  - open-source
  - multimodal
is_open_source: true
context_window: 8192
max_output: 8192
pricing:
  - image: 0
    input: 0.03
    output: 0.03
    provider: novita-ai
release_date: '2026-01-27'
capabilities:
  audio: false
  vision: true
  tool_use: false
  citations: false
  reasoning: false
---
Deepseek Deepseek-Ocr-2 is a Multimodal model.
OCR model for extracting structured text from documents and screenshots