# Qwen3-TTS-12Hz-0.6B-CustomVoice by Qwen: Open-Weight Model Source: https://savrn.com/models/qwen3-tts-12hz-0-6b-customvoice Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Runs On What it takes to serve Qwen3-TTS-12Hz-0.6B-CustomVoice (906M 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](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 8-bit | 0.9 GB | 1.1 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 4-bit | 0.5 GB | 0.5 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/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](https://savrn.com/ai-index/pricing/gpus), read Oct 23, 2026. [Qwen3-TTS-12Hz-0.6B-CustomVoice on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/qwen3-tts-12hz-0-6b-customvoice/gpus) ## SAVRN's Notes on Qwen3-TTS-12Hz-0.6B-CustomVoice A 906M-parameter voice model that Qwen ships as the CustomVoice variant of its 12 Hz TTS series: 9 premium timbres, plain-language style control and 3-second voice cloning, trained on over 5 million hours of speech across 10 languages per the card. Memory is 2.2 GB at 16-bit, 1.1 GB at 8-bit, 0.5 GB at 4-bit — the cheapest listed card, an MI300X at $1.85 an hour, is nowhere near the constraint. Concurrency is. Apache 2.0 puts the audio and any tuned voice into commercial products. The download line routes through ModelScope, and the 2.5 GB pull includes a separate speech_tokenizer checkpoint. Pin it against the technical report, arXiv 2601.15621; no context length or evaluations are on file, so run listening tests in every language you ship. At 995.5k downloads a month it is gaining fast, and no SAVRN Index host prices it by the token yet. SAVRN Research ## Model Card By Qwen, published under apache-2.0, revision 85e237c12c02. [Qwen3-TTS](https://github.com/QwenLM/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. - Paper: [Qwen3-TTS Technical Report](https://savrn.com/papers/qwen3-tts-technical-report) - GitHub: [QwenLM/Qwen3-TTS](https://github.com/QwenLM/Qwen3-TTS) - Demo: [Hugging Face Spaces](https://huggingface.co/spaces/Qwen/Qwen3-TTS) ### Key Features - Multilingual Synthesis: Supports Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, and Italian. - Intelligent Control: Adapts tone, rhythm, and emotional expression based on natural language instructions (e.g., "Speak in a very happy tone"). - Low Latency: Optimized for streaming generation with the Qwen3-TTS-Tokenizer-12Hz, achieving end-to-end synthesis latency as low as 97ms. ### Quickstart To use Qwen3-TTS, you can install the qwen-tts package: ``` pip install -U qwen-tts ``` #### Sample Usage [Read the full model card (333 words)](https://savrn.com/models/qwen3-tts-12hz-0-6b-customvoice/card) ## Configuration Architecture Qwen3TTSForConditionalGeneration Model type qwen3_tts ## Identity and Version Repository Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice Publisher Qwen Task Text to speech Modality Audio Library Not stated by the source Parameters 906M parameters Languages zh, en, ja, ko, de, fr, ru, pt Revision 85e237c12c027371202489a0ec509ded67b5e4b5 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 · 7.9 KB Tokenizer3 files · 4.5 MB Documentation1 file · 3.3 KB Repository1 file · 1.5 KB Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | model.safetensors | Weights | 1.8 GB | bc3c7e785eb9 | | speech_tokenizer/model.safetensors | Weights | 682.3 MB | 836b7b357f5e | | config.json | Configuration | 4.9 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 | 3.3 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](https://modelscope.cn/models/Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice) Released by Qwen through ModelScope. [Read the license](https://www.apache.org/licenses/LICENSE-2.0). ## Built From - Described by [arXiv:2601.15621](https://savrn.com/papers/qwen3-tts-technical-report) ## 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 - Derived from[Qwen3-TTS-12Hz-0.6B-CustomVoice-OpenVINO](https://savrn.com/models/qwen3-tts-12hz-0-6b-customvoice-openvino) ## Questions About Qwen3-TTS-12Hz-0.6B-CustomVoice ### How much GPU memory does Qwen3-TTS-12Hz-0.6B-CustomVoice need? About 2.2 GB at 16-bit and 0.5 GB at 4-bit: the weights (906M parameters) plus a working margin. A long context needs more. ### What is the cheapest GPU to run Qwen3-TTS-12Hz-0.6B-CustomVoice 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-CustomVoice commercially? Yes. Qwen3-TTS-12Hz-0.6B-CustomVoice 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 ### [qwen-tts-ngoc-huyen-voice](https://savrn.com/models/qwen-tts-ngoc-huyen-voice) [NGUYEN HUU TAM](https://savrn.com/model-publishers/huutamm1612) This Vietnamese text-to-speech model is fine-tuned from g-group-ai-lab/gwen-tts-0.6B, based on the Qwen3-TTS architecture. The exported package includes merged weights and the ngochuyen speaker in CustomVoice mode. No reference audio is required at inference time. The model was fine-tuned on approximately 90 hours of Ngọc Huyền's voice, collected by crawling YouTube and TikTok. Training used LoRA adaptation, adapter merging, and selective refinement of the talker and output head, followed by joint refinement with the code predictor. The final model was selected by the training pipeline and packaged with a speaker embedding derived from reference audio. If the repository is private… Open weights 906M parameters [View model](https://savrn.com/models/qwen-tts-ngoc-huyen-voice) Model · Text to speech ### [Qwen3-TTS-12Hz-0.6B-Base](https://savrn.com/models/qwen3-tts-12hz-0-6b-base) [Qwen](https://savrn.com/model-publishers/qwen) 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. To clone a voice and synthesize new content using the Base model, you can use the following code snippet: Qwen3-TTS covers 10 major languages (Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, and Italian) as well as multiple dialectal voice profiles to meet global application… Open weights apache-2.0 915M parameters [View model](https://savrn.com/models/qwen3-tts-12hz-0-6b-base) Model · Text to speech ### [indic-parler-tts](https://savrn.com/models/indic-parler-tts) [AI4Bharat](https://savrn.com/model-publishers/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](https://savrn.com/models/indic-parler-tts) Model · Text to speech ### [SemBridge](https://savrn.com/models/sembridge) [ASLP-lab](https://savrn.com/model-publishers/aslp-lab) SemBridge bridges discrete semantic supervision and continuous acoustic generation. A shared semantic tokenizer anchors both acoustic latents and autoregressive language-model states, improving linguistic fidelity while retaining continuous-latent generation. The unified model supports text-to-speech, note-conditioned singing, and mixed speech and singing. - [2026-10]: Released the inference code and model checkpoints. - [2026-08]: Released the arXiv paper. Use Python 3.10. The repository includes 10 complete examples from our Demo page: 4 TTS, 4 SVS, and 2 Unified cases, with reference audio and prepared requests. SVS and Unified include aligned lyrics, pitches, durations, and… Open weights apache-2.0 818M parameters 32,768 tokens [View model](https://savrn.com/models/sembridge) Model · Text to speech ### [VibeVoice-Realtime-0.5B](https://savrn.com/models/vibevoice-realtime-0-5b) [Microsoft](https://savrn.com/model-publishers/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](https://savrn.com/models/vibevoice-realtime-0-5b) Model · Text to speech ### [microsoft__VibeVoice-Realtime-0.5B](https://savrn.com/models/microsoft-vibevoice-realtime-0-5b) [Allinsama](https://savrn.com/model-publishers/allinsama) 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](https://savrn.com/models/microsoft-vibevoice-realtime-0-5b) ## Qwen [All models and datasets](https://savrn.com/model-publishers/qwen) ## Versions - [85e237c12c02](https://savrn.com/models/qwen3-tts-12hz-0-6b-customvoice/versions/85e237c12c02) · current 2026-10-18 ## Explore More - [All text to speech models](https://savrn.com/models/tasks/text-to-speech) - [All models under apache-2.0](https://savrn.com/models/licenses/apache-2-0) - [Model comparisons](https://savrn.com/models/comparisons) - [The model directory](https://savrn.com/models) - [Open model prices by host](https://savrn.com/ai-index/pricing/open-models) ## Source - Repository metadata, read 2026-10-19. - [Qwen website](https://qwen.ai/) - [Hugging Face record](https://huggingface.co/Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice) - [How the hub is built](https://savrn.com/model-hub/methodology)