llmrun
moondreamHfMoondream

Moondream3 Preview — Hardware Requirements & GPU Compatibility

Vision

Moondream 3 (Preview) is a 9.3-billion-parameter vision-language model from Moondream, built as a sparse Mixture-of-Experts with 64 experts per layer and only 2 billion parameters active per token. It targets visual reasoning such as object detection, counting, pointing, and visual question answering, aiming for accuracy closer to larger closed models while keeping inference light. All expert weights must fit in memory, but the active footprint keeps it usable on a single high-end consumer GPU once quantized. Moondream reports a usable context window of around 32,000 tokens. It ships under a Business Source License 1.1 with an additional use grant: free for personal, research, and most commercial use, but resale as a hosted service is blocked. Published in September 2025 as a preview, it introduces a mixture-of-experts design not present in earlier, dense Moondream models.

9.3B params · denseOther40.2K downloads694 likes287 quant downloads
B

Decent on a 24 GB GPU

Runs on24 GB BF16 · ~32 tok/s · 16KMacs from24 GB BF16 · ~11 tok/s

Other

Reference GPU · RTX 4090 24 GB

Estimates, ±20% · how we estimate

Run Moondream3 Preview

One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.

Recommended

vLLM
one command · needs vLLM installed
$vllm serve moondream/moondream3-preview --max-model-len 16384

vLLM repo with its own quantisation

Download: Weights on Hugging Face

install, OS, other quants

Ways to runOther ways

each opens step-by-step

Run in cloud

Runs on RTX 5090 (32 GB) (8 GB headroom) · BF16

Generation speed
~57 tok/s
generation speed
Cost per 1M output tokens
$2.42
per 1M output tokens
Deployon Vast.aiBest valuefrom $0.50/hr (opens in new tab)
Compare GPUs →
or

How Much VRAM Does Moondream3 Preview Need?

Select a quantization level; the listed hardware adheres to it.

QuantizationVRAM+32KOn RTX 4090 (reference)
20.4 GB26.1 GB16-bit · bfloat — Brain floating point 16 — preferred for trainingFits: 16K context · ~32 tok/s
est.calculated, no published file yet

Selected BF16

Brain floating point 16

23.1GBat 16K context

Weights
19.7 GB
KV cache
3.1 GB
Overhead
0.3 GB

Fits RTX 4090 (reference) · 0.9 GB free · ~32 tok/s

≥50 tok/s needs ≥1.57 TB/s · RTX 5090 ~57

Which Hardware Can Run Moondream3 Preview?

BF16 · 20.4 GB

Moondream3 Preview (BF16) requires 20.4 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 27+ GB is recommended. 7 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 122 devices with unified memory can run Moondream3 Preview, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Studio M1 Max (32 GB).

By VRAM tier · typical GPUs and Macs at BF16
8–16 GB · RTX 4060, RTX 3060 12GB, RTX 5060 Ti 16GBDoesn't fit — smallest quant needs 23.1 GB8 GB tier →12 GB tier →16 GB tier →
24 GB · RTX 4090referenceBF16 · 23.1 GB · ~32 tok/s · 16K24 GB tier →
Mac 32 GB · Mac Mini M4BF16 · 23.1 GB · ~4 tok/s · 32KMacs →
Mac 64 GB · Mac Studio M4 MaxBF16 · 23.1 GB · ~19 tok/s · 32KMacs →

Alternatives to Moondream3 Preview

Similar models

6
ModelScoreVRAMtok/sWhy it is similar
Qwen3.5 9BAlibaba99.46.6 GB~103same size
Granite Switch 4.1 8BIBM—8.4 GB~111same size
Ornith 1.5 9Bornith-ai—6.6 GB~103same size
Apertus V1.5 8Bswiss-ai—8.5 GB~111same size
MiniCPM O 4 5OpenBMB—8.3 GB~105same size
Qwen3 VL 8B InstructAlibaba—8 GB~112same size

Specifications

Publisher
moondream
Parameters
9.3B
Architecture
HfMoondream
Release Date
2025-09-11
License
Other

Resources

Frequently Asked Questions

How much VRAM does Moondream3 Preview need?

Moondream3 Preview requires 20.4 GB of VRAM at BF16.

VRAM = Weights + KV Cache + Overhead

Weights = 9.3B × 16 bits ÷ 8 = 18.5 GB

KV Cache + Overhead ≈ 1.9 GB (at 2K context + ~0.3 GB framework)

Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.

VRAM usage by quantization

20.4 GB

Learn more about VRAM estimation →

Can I run Moondream3 Preview on a Mac?

Yes — Mac Mini M5 Pro (24 GB) and 106 other Macs can run Moondream3 Preview. Apple Silicon uses unified memory, so the model shares RAM with the system. At BF16 you need at least 20.4 GB of usable unified memory (RAM minus macOS overhead).

Can I run Moondream3 Preview locally?

Yes — Moondream3 Preview can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.

How fast is Moondream3 Preview?

At BF16, Moondream3 Preview can reach ~235 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~32 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.

tok/s = (bandwidth GB/s ÷ model GB) × efficiency

Example: NVIDIA B200 → 8000 ÷ 20.4 × 0.65 = ~255 tok/s

Estimated speed at BF16 (20.4 GB)

~255 tok/s
~32 tok/s
~255 tok/s
~235 tok/s

Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.

Learn more about tok/s estimation →

What's the download size of Moondream3 Preview?

At BF16, the download is about 18.54 GB.

Which GPUs can run Moondream3 Preview?

7 consumer GPUs can run Moondream3 Preview at BF16. Top options include NVIDIA GeForce RTX 5090, AMD Radeon RX 7900 XTX, NVIDIA GeForce RTX 3090. 1 GPU has plenty of headroom for comfortable inference.

Which devices can run Moondream3 Preview?

122 devices with unified memory can run Moondream3 Preview at BF16 (20.4 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.