TAAFT
Free mode
100% free
Freemium
Free Trial
Deals
Companies
DeepSeek
Models
DeepSeek OCR 2

DeepSeek OCR 2

By DeepSeek
New Text Gen 7
Released: January 27, 2026

Overview

Second-generation DeepSeek OCR model, “Visual Causal Flow,” aimed at more human-like visual encoding, with dynamic resolution support and strong document-to-Markdown and layout-aware OCR for images and PDFs.

Description

DeepSeek-OCR-2 builds on DeepSeek-OCR with a “Visual Causal Flow” design that targets more human-like visual encoding and improved document understanding. It supports dynamic resolution, keeps vLLM and Transformers inference paths, and includes prompts for layout-aware document-to-Markdown conversion and lightweight “Free OCR” usage over images and PDFs.

About DeepSeek

DeepSeek is a Chinese AI firm specializing in large language models, based in Hangzhou.

Industry: Artificial Intelligence
Company Size: N/A
Location: Hangzhou, Zhejiang, CN
Website: https://deepseek.com
View Company Profile

Related Models

  • Composer 1 Gen 7
    Composer 1 is a controllable text-to-image model that turns long prompts and references into production-ready visuals with clean layout, readable small text, and precise regional edits.
    Text
    Released 3mo ago
  • Ling flash 2.0 Gen 7
    Ling-flash-2.0 is a high-speed multilingual instruction model built for very low latency and high throughput. It supports long context, tool and function calling, and clean JSON outputs, which makes it ideal for live chat, voice assistants, and real-time automation.
    Text
    Released 4mo ago
  • DeepSeek 3.2 Speciale Gen 7
    DeepSeek-V3.2-Speciale is a 685B-parameter research-only variant of DeepSeek-V3.2 that pushes open-weight reasoning ability to the limit, but disables tool calling and is intended purely for experimentation rather than everyday agent use.
    NewText
    Released 1mo ago
Last updated: January 30, 2026
0 AIs selected
Clear selection
#
Name
Task