Dolphin-v2

ByteDance/Dolphin-v2

Image Text to Text3.8B parameterstransformers
Tracked by OpenModelStats since September 8, 2026View on Hugging Face ↗
Downloads 30D
2,155
Total downloads
23.6K
Gained 7D
+1,545(increase)
7.2%(up) vs baseline
Gained 30D
—
Likes
115
+0 in 7D
Rank · tracked models
#20,783
804(down 804 places)this week
Parameters
3.8B
Last source update
Dec 12, 2025
Momentum
64.1
  • Dolphin-v2 gained 1,545 downloads during the last seven tracked days.
  • It moved from #1955 to #2020 among tracked Image Text to Text models this week.

Downloads over time

Daily gain between consecutive observations

Likes over time

Rank among tracked models

Lower is better

Overview

Dolphin-v2 is a image text to text model published by ByteDance. OpenModelStats has tracked the model since Sep 8, 2026, most recently observing it 8h ago.

Published
Dec 1, 2025
Last updated
Dec 12, 2025
Library
transformers
License
—
Task
Image Text to Text
Parameters
3,754,622,976
Spaces
—
Derivative models
1
Velocity 7D/day
221
transformerssafetensorsqwen2_5_vlimage-text-to-textdocument-parsingdocument-understandingdocument-intelligenceocrlayout-analysistable-extractionformula-recognitioncode-extraction

Model family

Based on base-model relationships reported by the source's metadata.

Derivative models

Similar models

Models similar to Dolphin-v2
ModelDL 30D
Qwen2.5-VL-3B-InstructQwen2.3M
Qwen2.5-VL-3B-Instruct-AWQQwen334.2K
Nanonets-OCR2-3Bnanonets268.7K
Nanonets-OCR-snanonets58.3K
typhoon-ocr-3btyphoon-ai29.4K
churro-3Bstanford-oval23.8K
ByteDance/Dolphin-v2 — Downloads, Growth & Stats | OpenModelStats