other
1.9B parameters
Commercial OK
Multimodal
Reviewed May 2026Moondream 2
Tiny vision-language model. ~1.9B; designed for edge / embedded multimodal use cases. Apache 2.0.
License: Apache 2.0·Released Jul 22, 2024·Context: 2,048 tokens
Overview
Tiny vision-language model. ~1.9B; designed for edge / embedded multimodal use cases. Apache 2.0.
Strengths
- Apache 2.0 multimodal at 1.9B
- Edge-deployable
Weaknesses
- Quality ceiling at 1.9B parameters
Quantization variants
Each quantization trades model quality for file size and VRAM. Q4_K_M is the most popular starting point.
| Quantization | File size | VRAM required |
|---|---|---|
| Q4_K_M | 1.2 GB | 2 GB |
Get the model
HuggingFace
Original weights
huggingface.co/vikhyatk/moondream2
Source repository — direct quantization required.
Hardware that runs this
Cards with enough VRAM for at least one quantization of Moondream 2.
Frequently asked
What's the minimum VRAM to run Moondream 2?
2GB of VRAM is enough to run Moondream 2 at the Q4_K_M quantization (file size 1.2 GB). Higher-quality quantizations need more.
Can I use Moondream 2 commercially?
Yes — Moondream 2 ships under the Apache 2.0, which permits commercial use. Always read the license text before deployment.
What's the context length of Moondream 2?
Moondream 2 supports a context window of 2,048 tokens (about 2K).
Does Moondream 2 support images?
Yes — Moondream 2 is multimodal and accepts text + vision inputs. Vision support requires a runner that handles its image-conditioning architecture.
Source: huggingface.co/vikhyatk/moondream2
Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify model claims.
Related — keep moving
Compare hardware
Buyer guides
Before you buy
Verify Moondream 2 runs on your specific hardware before committing money.