MO

moonshine-base

MoonshineGGUF by cstrUpdated 2026-08-02
Download on Hugging Face
Parameters62M
Context—
Layers—
Architecturemoonshine
Smallest quant0.0 GB
Quick answer

moonshine-base needs about 0.6 GB of memory at Q8_0 (a 0.1 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 3 quants, from 0.0 GB (Q4_K) to 0.2 GB.

Moonshine Base – 61M‑parameter English ASR model in GGUF format

Moonshine Base is a 61M‑parameter sequence‑to‑sequence ASR model trained for English speech transcription. It is available in GGUF format and runs on CPU only. The model is released under the MIT license.

Good at

  • Real‑time English speech transcription
  • Low‑memory deployment
  • CPU‑only inference

Limits

  • No vision support
  • Only English language
  • Requires tokenizer.bin in same folder

Best for

  • Deploying on edge devices
  • Integrating with CrispASR
  • Building low‑cost speech apps

Who it’s for. AI developers who need an English ASR solution that fits in tight memory budgets and can run on CPUs. The model is suitable for real‑time transcription on low‑cost hardware.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_K0.0 GB 0.6 GB
Fits Get file →
Q8_00.1 GB 0.6 GB
Fits Get file →
—0.2 GB 0.8 GB
Fits Get file →

For 8 GB VRAM, the Q4_K variant (0.05 GB file, 0.6 GB memory at 8k context) is ideal. If more memory is available, Q8_0 (0.08 GB file, 0.6 GB memory) offers higher precision. The unknown quant (0.25 GB file, 0.8 GB memory) is the largest option.

What GPU do you need for moonshine-base?

These figures add the file size, an fp16 KV cache at an 8K context window and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is the main file (0.2 GB file, about 0.8 GB in use).

Which moonshine-base quant should you choose?

Q8_0 (0.1 GB) is the middle option in this repository. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.

If memory is tight, the smallest file is Q4_K at 0.0 GB, with the quality trade-offs that come with fewer bits. The usual sweet spot between file size and quality, and the quant most people start with.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturemoonshine
Parameters62M

Source

Quantized bycstr
FamilyMoonshine
Licensemit
Downloads (HF)962
Last updated2026-08-02

Frequently asked questions about moonshine-base

How much VRAM does moonshine-base need?

About 0.6 GB at Q8_0 with an 8K context window. That is the 0.1 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run moonshine-base on an 8 GB GPU?

Yes, with the smaller files: the 0.2 GB file needs about 0.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of moonshine-base GGUF?

It depends on the quant: Q4_K is 0.0 GB, Q8_0 is 0.1 GB, file is 0.2 GB.

Where can I download moonshine-base GGUF?

The files are on Hugging Face in the cstr/moonshine-base-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does moonshine-base use?

Hugging Face lists the license as mit. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run Moonshine Base on 8 GB VRAM?

Yes. The Q4_K and Q8_0 variants use only 0.6 GB of VRAM at 8k context, well within 8 GB.

Does Moonshine Base support GPU inference?

No. The model is designed for CPU inference; GPU is not required.

What languages does Moonshine Base support?

Only English. The model was trained and evaluated on English ASR tasks.

Other GGUFs of moonshine-base

Similar models

Browse all →

More from cstr

All by cstr →

More Moonshine models

All Moonshine →

Other tiny models

All tiny models →