Name: Qwen3 TTS Tokenizer 12Hz
Author: Qwen

Overview

The Qwen3 TTS Tokenizer 12Hz is a specialized audio-to-audio component designed to bridge the gap between raw waveforms and discrete token representations. Operating at a 12Hz sampling rate for tokens, it focuses on efficient compression of acoustic features while preserving the phonetic and prosodic integrity required for high-fidelity speech synthesis. For developers, this means a significant reduction in sequence length compared to raw audio processing, enabling faster inference and lower memory overhead when integrating speech capabilities into LLM pipelines. It is particularly suited for building low-latency voice assistants or multimodal agents where seamless audio-in/audio-out loops are critical. Licensed under Apache-2.0, it offers a flexible, open-standard approach to audio tokenization that integrates easily into existing PyTorch or JAX workflows.

Highlights

Efficient 12Hz audio tokenization for reduced sequence length
Optimized for low-latency audio-to-audio multimodal pipelines
Preserves critical prosodic features for natural speech synthesis
Permissive Apache-2.0 license for commercial and open integration

Usage

Install

# Install Hugging Face transformers
pip install transformers torch

SDK Usage

# Load model with transformers
from transformers import AutoModel, AutoTokenizer

model = AutoModel.from_pretrained("Qwen/Qwen3-TTS-Tokenizer-12Hz")
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3-TTS-Tokenizer-12Hz")

Hugging Face Download

We recommend downloading the model via the Hugging Face CLI or Hub SDK.

Guidance：Before downloading, install huggingface_hub with:

Guidance

pip install -U huggingface_hub

CLI Download

Download the full repository

huggingface-cli download Qwen/Qwen3-TTS-Tokenizer-12Hz

Download a single file to a local folder (e.g. config.json into ./dir)

huggingface-cli download Qwen/Qwen3-TTS-Tokenizer-12Hz config.json --local-dir ./dir

See the official docs for more CLI options

SDK Download

# 模型下载
from huggingface_hub import snapshot_download
model_dir = snapshot_download('Qwen/Qwen3-TTS-Tokenizer-12Hz')

Git Download

Make sure git-lfs is installed first

Git Download

git lfs install
git clone https://huggingface.co/Qwen/Qwen3-TTS-Tokenizer-12Hz

To skip LFS large-file downloads, use:

Skip LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/Qwen/Qwen3-TTS-Tokenizer-12Hz

Model files are hosted on the Hugging Face Hub — download directly via HF CLI / SDK / Git, not through this site.

PyTorch / Transformers Usage

Install Transformers

pip install -U transformers torch

Load the model and run inference

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained('Qwen/Qwen3-TTS-Tokenizer-12Hz')
tokenizer = AutoTokenizer.from_pretrained('Qwen/Qwen3-TTS-Tokenizer-12Hz')

Model Download

We recommend downloading the model via the ModelScope CLI or SDK.

Guidance：Before downloading, install ModelScope with:

Guidance

pip install modelscope

CLI Download

Download the full repository

modelscope download --model Qwen/Qwen3-TTS-Tokenizer-12Hz

Download a single file to a local folder (e.g. README.md into ./dir)

modelscope download --model Qwen/Qwen3-TTS-Tokenizer-12Hz README.md --local_dir ./dir

See the docs for more CLI options

SDK Download

# 模型下载
from modelscope import snapshot_download
model_dir = snapshot_download('Qwen/Qwen3-TTS-Tokenizer-12Hz')

Git Download

Make sure git-lfs is installed first

Git Download

git lfs install
git clone https://www.modelscope.cn/Qwen/Qwen3-TTS-Tokenizer-12Hz.git

To skip LFS large-file downloads, use:

Skip LFS

GIT_LFS_SKIP_SMUDGE=1 git clone https://www.modelscope.cn/Qwen/Qwen3-TTS-Tokenizer-12Hz.git

ModelScope 模型页直接下载模型文件；无需将模型文件放在本站服务器。

Notebook Quickstart

Install the ModelScope library

pip install "modelscope[audio,cv,nlp,multi-modal,science]" -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html

Load the model and run inference

from modelscope.pipelines import pipeline
from modelscope.utils.constant import Tasks

p = pipeline('text-generation', 'Qwen/Qwen3-TTS-Tokenizer-12Hz')

Full Documentation

来源: HuggingFace

---
license: apache-2.0
pipeline_tag: audio-to-audio
tags:

audio

tts

speech

codec

---
---

Qwen3-TTS-Tokenizer-12Hz

This repository contains the Qwen3-TTS-Tokenizer-12Hz, as presented in the paper Qwen3-TTS Technical Report.

Qwen3-TTS-Tokenizer-12Hz achieves extreme bitrate reduction and ultra-low-latency streaming, enabling immediate first-packet emission through its 12.5 Hz, 16-layer multi-codebook design and a lightweight causal ConvNet.

Paper: Qwen3-TTS Technical Report

GitHub Repository: QwenLM/Qwen3-TTS

Demo: Hugging Face Space

Quickstart

Environment Setup

Install the qwen-tts Python package from PyPI:

bash

pip install -U qwen-tts

Tokenizer Encode and Decode

You can encode audio into discrete tokens for storage or transport and decode them back into speech using the snippet below:

python

import soundfile as sf
from qwen_tts import Qwen3TTSTokenizer
tokenizer = Qwen3TTSTokenizer.from_pretrained(
    "Qwen/Qwen3-TTS-Tokenizer-12Hz",
    device_map="cuda:0",
)
Encode audio from a URL (or local path)
enc = tokenizer.encode("https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen3-TTS-Repo/tokenizer_demo_1.wav")
Decode codes back into waveforms
wavs, sr = tokenizer.decode(enc)
sf.write("decode_output.wav", wavs[0], sr)

Overview

Introduction

Qwen3-TTS covers 10 major languages (Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, and Italian) as well as multiple dialectal voice profiles. Key features:

Powerful Speech Representation: Powered by the self-developed Qwen3-TTS-Tokenizer-12Hz, it achieves efficient acoustic compression and high-dimensional semantic modeling of speech signals. It fully preserves paralinguistic information and acoustic environmental features.

Extreme Low-Latency Streaming Generation: Based on the innovative Dual-Track hybrid streaming generation architecture, it can output the first audio packet immediately after a single character is input, with end-to-end synthesis latency as low as 97ms.

Model Architecture

Released Tokenizers

| Tokenizer Name | Description |
|----------------|-------------|
| Qwen3-TTS-Tokenizer-12Hz | The Qwen3-TTS-Tokenizer-12Hz model which can encode the input speech into codes and decode them back into speech. |

Evaluation

For detailed evaluation results on speech generation consistency, speaker similarity, and tokenizer benchmarks (ASR tasks, PESQ, STOI, UTMOS), please refer to the technical report or the GitHub repository.

Citation

bibtex

@article{Qwen3-TTS,
  title={Qwen3-TTS Technical Report},
  author={Hangrui Hu and Xinfa Zhu and Ting He and Dake Guo and Bin Zhang and Xiong Wang and Zhifang Guo and Ziyue Jiang and Hongkun Hao and Zishan Guo and Xinyu Zhang and Pei Zhang and Baosong Yang and Jin Xu and Jingren Zhou and Junyang Lin},
  journal={arXiv preprint arXiv:2601.15621},
  year={2026}
}