Deepseek OCR

Parameters

3B

Published

Oct 21, 2025

Deepseek OCR is an AI model developed by DeepSeek (China), first published in October 2025. It works in the vision and language domain, on tasks such as character recognition (ocr) and visual question answering.

Epoch AI has no training-compute estimate for this model. The model has 3,000,000,000 parameters. Training ran on 160 NVIDIA A100 SXM4 40 GB.

Access: Open weights (unrestricted). Its weights are openly available. It is built on top of Segment Anything Model,CLIP (ViT L/14@336px). Epoch AI rates the confidence of this record as confident.

Full record

Organization: DeepSeek
Country of organization: China
Domain: Vision, Language
Task: Character recognition (OCR), Visual question answering
Parameters: 3,000,000,000
Training hardware: NVIDIA A100 SXM4 40 GB
Chips used: 160
Training power draw: 125.0 kW
Model accessibility: Open weights (unrestricted)
Open weights: Yes
Base model: Segment Anything Model, CLIP (ViT L/14@336px)
Epoch confidence: Confident