FU

Fun-CosyVoice3-0.5B-2512-voices

FunGGUF by cstrUpdated 2026-08-05
Download on Hugging Face
Parameters7M
Context—
Layers—
Architecturecosyvoice3-voices
Smallest quant0.0 GB
Quick answer

Fun-CosyVoice3-0.5B-2512-voices needs about 0.9 GB of memory at Q4_K (a 0.4 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 11 quants, from 0.0 GB to 1.3 GB (F16).

Fun-CosyVoice3-0.5B-2512: 0.5B multilingual TTS with zero‑shot voice cloning

Fun-CosyVoice3-0.5B-2512 is a text‑to‑speech model that supports nine languages and 18 Chinese dialects. It uses a three‑stage pipeline and offers streaming inference with low latency. The model is licensed under Apache‑2.0.

Good at

  • Zero‑shot multilingual voice cloning
  • Low‑latency streaming inference
  • Pronunciation inpainting for Chinese Pinyin and English CMU phonemes

Limits

  • No vision support
  • Requires a dedicated GGUF file for each quant
  • Large memory for higher‑quality quantizations

Best for

  • Voice‑over production in multiple languages
  • Real‑time TTS in chat applications
  • Audio‑based accessibility tools

Who it’s for. Developers building multilingual TTS services or voice‑over pipelines. The model is suitable for applications that need zero‑shot voice cloning and low‑latency streaming. It is ideal for teams that can manage multiple GGUF files and have a suitable GPU or CPU environment.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.0 GB 0.5 GB
Fits Get file →
F160.0 GB 0.5 GB
Fits Get file →
F160.0 GB 0.5 GB
Fits Get file →
Q4_K0.1 GB 0.7 GB
Fits Get file →
Q8_00.4 GB 0.9 GB
Fits Get file →
Q4_K0.4 GB 0.9 GB
Fits Get file →
Q4_K0.4 GB 0.9 GB
Fits Get file →
F160.5 GB 1.0 GB
Fits Get file →
F160.7 GB 1.2 GB
Fits Get file →
F161.3 GB 1.9 GB
Fits Get file →
F161.3 GB 1.9 GB
Fits Get file →

For 8 GB VRAM, use the Q4_K quant with a 0.15 GB file; it needs about 0.7 GB memory at 8 k context. If more VRAM is available, the Q8_0 quant (0.36 GB file, 0.9 GB memory) offers higher quality. The F16 quant (0.48 GB file, 1 GB memory) is a good balance for 12 GB VRAM setups.

What GPU do you need for Fun-CosyVoice3-0.5B-2512-voices?

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 F16 (1.3 GB file, about 1.9 GB in use).

Which Fun-CosyVoice3-0.5B-2512-voices quant should you choose?

Q4_K (0.4 GB) is the middle option in this repository. The usual sweet spot between file size and quality, and the quant most people start with.

If memory is tight, the smallest file is the first one in the table at 0.0 GB, with the quality trade-offs that come with fewer bits.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturecosyvoice3-voices
Parameters7M

Source

Quantized bycstr
FamilyFun
Licenseapache-2.0
Downloads (HF)12,909
Last updated2026-08-05

Frequently asked questions about Fun-CosyVoice3-0.5B-2512-voices

How much VRAM does Fun-CosyVoice3-0.5B-2512-voices need?

About 0.9 GB at Q4_K with an 8K context window. That is the 0.4 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 Fun-CosyVoice3-0.5B-2512-voices on an 8 GB GPU?

Yes, with F16: the 1.3 GB file needs about 1.9 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Fun-CosyVoice3-0.5B-2512-voices GGUF?

It depends on the quant: file is 0.0 GB, F16 is 0.0 GB, F16 is 0.0 GB, Q4_K is 0.1 GB, Q8_0 is 0.4 GB, Q4_K is 0.4 GB and 5 more.

Where can I download Fun-CosyVoice3-0.5B-2512-voices GGUF?

The files are on Hugging Face in the cstr/cosyvoice3-0.5b-2512-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 Fun-CosyVoice3-0.5B-2512-voices use?

Hugging Face lists the license as apache-2.0. 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 Fun-CosyVoice3-0.5B-2512 on 8 GB VRAM?

Yes, using the Q4_K quant (0.15 GB file) which requires approximately 0.7 GB memory at 8 k context.

Does the model support voice cloning?

Yes, it supports zero‑shot voice cloning across nine languages and 18 Chinese dialects.

What license does the model use?

The model is released under the Apache‑2.0 license.

Other GGUFs of Fun-CosyVoice3-0.5B-2512

Similar models

Browse all →

More from cstr

All by cstr →

More Fun models

All Fun →

Other tiny models

All tiny models →