Models
GitHub
Discord
Docs
Cloud
Sign in
Download
Models
Download
GitHub
Discord
Docs
Cloud
Sign in
deepseek-ocr
4,667
Downloads
Updated
yesterday
DeepSeek-OCR is a vision-language model that can perform token-efficient OCR.
DeepSeek-OCR is a vision-language model that can perform token-efficient OCR.
Cancel
vision
3b
Name
3 models
Size
Context
Input
deepseek-ocr:latest
0e7b018b8a22
• 6.7GB • 8K context window •
Text, Image input • yesterday
Text, Image input • yesterday
deepseek-ocr:latest
6.7GB
8K
Text, Image
0e7b018b8a22
· yesterday
deepseek-ocr:3b
latest
0e7b018b8a22
• 6.7GB • 8K context window •
Text, Image input • yesterday
Text, Image input • yesterday
deepseek-ocr:3b
latest
6.7GB
8K
Text, Image
0e7b018b8a22
· yesterday
deepseek-ocr:3b-bf16
0e7b018b8a22
• 6.7GB • 8K context window •
Text, Image input • yesterday
Text, Image input • yesterday
deepseek-ocr:3b-bf16
6.7GB
8K
Text, Image
0e7b018b8a22
· yesterday