Moondream 3 Preview
Moondream 3 Preview is a compact frontier-oriented vision-language model built for fast visual reasoning, grounding, OCR, object detection, pointing, and structured output. It uses a 9B MoE architecture with 2B active parameters and extends context length to 32K, aiming to deliver strong real-world vision performance while staying efficient and inexpensive to run.
- Type
- Multimodal
- Weights
- Open
- Downloads
- 11k
- Likes
- 685
- Pipeline
- image text to text
- License
- other
Model hub
moondream/moondream3-preview
- Repo
- moondream/moondream3-preview
- Library
- transformers
- Access
- Public
Technical tags and languages
More from Moondream
6
-
25 DL
Moondream 0.5B
Image
Moondream 0.5B is a tiny open-source vision-language model built for edge devices and mobile platforms. With…
-
Moondream 1
Multimodal
Small, efficient open-source vision-language model designed to run broadly on many devices.
-
moondream 2
Multimodal
Moondream 2 is a small open vision-language model designed to run efficiently across many environments. It is…
-
Photon
Multimodal
Photon is Moondream’s real-time vision-language model aimed at production video and image analysis. It is des…
About Moondream
1
State-of-the-art visual understanding at speeds that make continuous processing possible. Point, detect, count, and reason—without compromise.