Model reference · open weights
qwen3-tts-f16 is an open-weight audio or speech model from OpenVoiceOS. qwen3-tts-0.6b-f16 (BF16) weighs 2.2 GB; the smallest configuration that runs it is RTX 3060 12 GB.
For self-hosting, compare this build's requirements below with the GPU you plan to rent.
Summary of the OpenVoiceOS/qwen3-tts-0.6b-f16 model card, 2026-10-01
What it is
| Released by | OpenVoiceOS |
|---|---|
| Released | 2026-03-21 |
| VRAM | 2.2 GB for the weights |
What it runs on
| Card | One stream | Memory |
|---|---|---|
| RTX 3060 12 GB | fits | 11.6 GB |
| RTX 4060 Ti 16 GB | fits | 15.4 GB |
| RTX 3090 24 GB | fits | 23.4 GB |
| RTX 4090 24 GB | fits | 23.4 GB |
| RTX 5090 32 GB | fits | 31.0 GB |
| L40S 48 GB | fits | 44.0 GB |
| A100 80 GB | fits | 78.2 GB |
| H100 80 GB | fits | 78.1 GB |
| RTX PRO 6000 Blackwell 96 GB | fits | 93.8 GB |
| DGX Spark (GB10) 128 GB unified | fits | 107 GB |
| H200 141 GB | fits | 138 GB |
| B200 180 GB | fits | 176 GB |
From the model card
This model is an quantized gguf-format export of Qwen/Qwen3-TTS-12Hz-0.6B-Base for ease of use in edge devices and CPU-based inference environments. The original model is transformed into gguf with F16 tensors by the script convert_hf_to_gguf.py and then further quantized, if needed, using the tool quantize from the same repo.
The requirements can be installed as
$ pip install git+https://github.com/femelo/py-qwen3-tts-cpp
from py_qwen3_tts_cpp.model import Qwen3TTSModel
# Initialize the model (it handles downloading from this repo)
model = Qwen3TTSModel(
tts_model="qwen3-tts-0.6b-f16",
n_threads=4
)
# Synthesize from text
result = model.synthesize("Hello, world!", language="en")
model.save_audio(result, "output.wav")
Refer to https://github.com/femelo/py-qwen3-tts-cpp for more details.
The license is derived from the original model: Apache 2.0. For more details, please refer to Qwen/Qwen3-TTS-12Hz-0.6B-Base.
Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.
How it works