DeepSeek-OCR-2

TitanML/DeepSeek-OCR-2

Image Text to Text3.4B parameterstransformersLicense: apache-2.0
Tracked by OpenModelStats since September 3, 2026View on Hugging Face ↗
Downloads 30D
442
Total downloads
24.6K
Gained 7D
+107(increase)
0.4%(up) vs baseline
Gained 30D
—
Likes
0
+0 in 7D
Rank · tracked models
#39,872
4279(down 4279 places)this week
Parameters
3.4B
Last source update
Aug 21, 2026
Momentum
32.1
  • DeepSeek-OCR-2 gained 107 downloads during the last seven tracked days.
  • It moved from #3382 to #3815 among tracked Image Text to Text models this week.

Downloads over time

Rolling recent-window downloads as reported

Likes over time

Rank among tracked models

Lower is better

Overview

DeepSeek-OCR-2 is a image text to text model published by TitanML. OpenModelStats has tracked the model since Sep 3, 2026, most recently observing it Oct 3, 2026.

Published
Apr 6, 2026
Last updated
Aug 21, 2026
Library
transformers
License
apache-2.0
Task
Image Text to Text
Parameters
3,389,119,360
Spaces
—
Derivative models
—
Velocity 7D/day
15
transformerssafetensorsdeepseek_vl_v2feature-extractiondeepseekvision-languageocrcustom_codeimage-text-to-textconversationalmultilingualarxiv:2601.20552

Similar models

Models similar to DeepSeek-OCR-2
ModelDL 30D
DeepSeek-OCR-2deepseek-ai794K
DeepSeek-OCR-2deepseek-community33.4K
DeepSeek-OCR-2unsloth6,047
DeepSeek-OCR-2-6bitmlx-community923
DeepSeek-OCR-2-GPTQ-INT8bartsolutions656
DeepSeek-OCR-2-8bitmlx-community655
TitanML/DeepSeek-OCR-2 — Downloads, Growth & Stats | OpenModelStats