Qwen3-TTS-12Hz-0.6B-Base by Qwen: Open-Weight Model

Runs On

What it takes to serve Qwen3-TTS-12Hz-0.6B-Base (915M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.

Precision	Weights	Memory needed	Cheapest setup	Per hour	Also fits
16-bit	1.8 GB	2.2 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00
8-bit	0.9 GB	1.1 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00
4-bit	0.5 GB	0.5 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00

Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.

SAVRN's Notes on Qwen3-TTS-12Hz-0.6B-Base

A voice clone from three seconds of audio is what the Base checkpoint is built for, and it does the job in 2.2 GB of memory at 16-bit, with 1.8 GB of weights. The cheapest listed setup is one MI300X with 192 GB at $1.85 an hour on demand, so nobody buys a card for this; you run it on whatever is already in the rack. Ten languages are covered, from more than 5 million hours of speech.

Apache 2.0 makes a voice product legal to ship: commercial use, modification and redistribution are allowed, with notices kept. Check the shape of the release. The file counts 915M parameters despite the 0.6B in the name, no context length is listed, the library field is empty and safetensors is the only format, so confirm your serving stack loads the Qwen3TTSForConditionalGeneration architecture. The technical report at arXiv:2601.15621 is the only paper attached.

SAVRN Research, 2026-09-18

Model Card

By Qwen, published under apache-2.0, revision 5d83992436ea.

Qwen3-TTS Technical Report | GitHub Repository | Hugging Face Demo

Qwen3-TTS is a family of advanced multilingual, controllable, robust, and streaming text-to-speech models. Trained on over 5 million hours of speech data spanning 10 languages, Qwen3-TTS supports state-of-the-art 3-second voice cloning and description-based control.

This specific checkpoint is the 0.6B Base model, which is capable of rapid voice cloning from a user-provided audio input.

Quickstart

Installation

pip install -U qwen-tts
# Optional: for optimized performance
pip install -U flash-attn --no-build-isolation

Sample Usage (Voice Clone)

To clone a voice and synthesize new content using the Base model, you can use the following code snippet:

Read the full model card (380 words)

Configuration

Architecture: Qwen3TTSForConditionalGeneration
Model type: qwen3_tts

Identity and Version

Repository: Qwen/Qwen3-TTS-12Hz-0.6B-Base
Publisher: Qwen
Task: Text to speech
Modality: Audio
Library: Not stated by the source
Parameters: 915M parameters
Languages: zh, en, ja, ko, de, fr, ru, pt
Revision: 5d83992436eae1d760afd27aff78a71d676296fc
First published: 2026-01-21
Last updated: 2026-01-29

Files and Weights

13 files, 2.5 GB in total. The weights are 2 files totalling 2.5 GB in safetensors.

Weights2 files · 2.5 GB

Configuration6 files · 7.5 KB

Tokenizer3 files · 4.5 MB

Documentation1 file · 3.6 KB

Repository1 file · 1.5 KB

Every file

File	Type	Size	SHA-256
model.safetensors	Weights	1.8 GB	180b3b10eb1c
speech_tokenizer/model.safetensors	Weights	682.3 MB	836b7b357f5e
config.json	Configuration	4.5 KB	—
generation_config.json	Configuration	245 B	—
preprocessor_config.json	Configuration	127 B	—
speech_tokenizer/config.json	Configuration	2.3 KB	—
speech_tokenizer/configuration.json	Configuration	76 B	—
speech_tokenizer/preprocessor_config.json	Configuration	234 B	—
README.md	Documentation	3.6 KB	—
.gitattributes	Repository	1.5 KB	—
merges.txt	Tokenizer	1.7 MB	—
tokenizer_config.json	Tokenizer	7.3 KB	—
vocab.json	Tokenizer	2.8 MB	—

License and Download

License: apache-2.0
Access: Open weights, no gate
Download size: 2.5 GB

Download from Qwen

Released by Qwen through ModelScope. Read the license.

Built From

Described by arXiv:2601.15621

Memory Requirements

Precision	Weights in memory
As published	2.5 GB
16-bit	1.8 GB
8-bit	0.9 GB
4-bit	0.5 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Built on This Model

Quantized fromQwen3-TTS-GGUF
Derived fromQwen3-TTS-GGUF

Questions About Qwen3-TTS-12Hz-0.6B-Base

How much GPU memory does Qwen3-TTS-12Hz-0.6B-Base need?

About 2.2 GB at 16-bit and 0.5 GB at 4-bit: the weights (915M parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run Qwen3-TTS-12Hz-0.6B-Base on?

At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.

Can I use Qwen3-TTS-12Hz-0.6B-Base commercially?

Yes. Qwen3-TTS-12Hz-0.6B-Base is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.

Similar Models

Model · Text to speech

Qwen3-TTS-12Hz-0.6B-CustomVoice

Qwen

Qwen3-TTS is a series of advanced multilingual, controllable, robust, and streaming text-to-speech models developed by the Qwen team. This specific checkpoint is the 0.6B CustomVoice variant, based on the 12Hz tokenizer. It supports 9 premium timbres and allows for fine-grained style control over target voices via natural language instructions across 10 major languages. To use Qwen3-TTS, you can install the qwen-tts package: For Qwen3-TTS-12Hz-0.6B-CustomVoice, the following speakers are supported. We recommend using each speaker’s native language for the best results: If you find Qwen3-TTS useful for your research, please consider citing

Open weights apache-2.0 906M parameters

View model

Model · Text to speech

indic-parler-tts

AI4Bharat

Indic Parler-TTS is a multilingual Indic extension of Parler-TTS Mini. It is a fine-tuned version of Indic Parler-TTS Pretrained, trained on a 1,806 hours multilingual Indic and English dataset. Indic Parler-TTS Mini can officially speak in 20 Indic languages, making it comprehensive for regional language technologies, and in English. The 21 languages supported are: Assamese, Bengali, Bodo, Dogri, English, Gujarati, Hindi, Kannada, Konkani, Maithili, Malayalam, Manipuri, Marathi, Nepali, Odia, Sanskrit, Santali, Sindhi, Tamil, Telugu, and Urdu. Thanks to its better prompt tokenizer, it can easily be extended to other languages. This tokenizer has a larger vocabulary and handles byte…

Access requested at publisher apache-2.0 938M parameters transformers

View model

Model · Text to speech

VibeVoice-Realtime-0.5B

Microsoft

VibeVoice-Realtime is a lightweight real‑time text-to-speech model supporting streaming text input and robust long-form speech generation. It can be used to build realtime TTS services, narrate live data streams, and let different LLMs start speaking from their very first tokens (plug in your preferred model) long before a full answer is generated. It produces initial audible speech in ~300 ms (hardware dependent). ▶ Watch demo video (Launch your own realtime demo via the websocket example in Usage) Although the model is primarily built for English, we found that it still exhibits a certain level of multilingual capability—and even performs reasonably well in some languages. We provide nine…

Open weights mit 1B parameters transformers

View model

Model · Text to speech

OmniVoice

K2 FSA

OmniVoice is a massively multilingual zero-shot text-to-speech (TTS) model supporting over 600 languages. Built on a novel diffusion language model-style architecture, it delivers high-quality speech with superior inference speed, supporting voice cloning and voice design. - 600+ Languages Supported: The broadest language coverage among zero-shot TTS models. To get started, install the omnivoice library: You can use OmniVoice for zero-shot voice cloning as follows: For more generation modes (e.g., voice design), functions (e.g., non-verbal symbols, pronunciation correction) and comprehensive usage instructions, see our GitHub Repository. You can directly discuss on GitHub Issues. You can…

Open weights 613M parameters 40,960 tokens omnivoice

View model

Model · Text to speech

gepard-1.0

NineNineSix

GEnerative, Prosody-aware, Autoregressive text-to-speech model for Realtime Dialogue Gepard is a text-to-speech model built for real-time conversation. It starts speaking the moment text begins arriving, generating audio piece by piece instead of waiting for a full sentence — so it feels like a live voice, not a recording. It's a single language model that learned text and speech together, so the output carries natural rhythm and timing rather than the flat, stitched tone of older pipelines. The name evokes "Gepard"(/geh-PART/), German for cheetah — a nod to the model's low-latency, high-throughput streaming. - One clean pass per frame — the whole audio frame (32 orthogonal FSQ channels) is…

Open weights apache-2.0 556M parameters 262,144 tokens transformers

View model

Model · Text to speech

VieNeu-TTS-v2

Pham Nguyen Ngoc Bao

VieNeu-TTS-v2 is the next generation of Vietnamese TTS, designed for Natural Communication, Podcasts, and Bilingual (En-Vi) Code-switching. This project features the flagship VieNeu-TTS-v2 architecture: Tác giả: Phạm Nguyễn Ngọc Bảo Training high-quality TTS models requires significant GPU resources. If you find this model useful, please consider supporting the development: Install the SDK to integrate VieNeu-TTS-0.3B into your research or applications: Deploy VieNeu-TTS as a high-performance API Server (powered by LMDeploy) with a single command. Start the Server with a Public Tunnel (No port forwarding needed): Once the server is running, you can connect from anywhere (Colab, Web Apps…

Open weights apache-2.0 294M parameters 4,096 tokens

View model