Qwen3-TTS
Open voices you can design by description or clone in 3 seconds.
FreeOpen sourceRuns on your PC
- Best for
- Free voices on a gaming PC
- On your PC
- Technical
- Works on
- Windows, Mac, Linux
- License
- Apache-2.0
What’s good
- Describe a voice in words and it creates it
- Clones a voice from 3 seconds of audio
- Apache license, 10 major languages
Downsides
- Needs Python and ideally an NVIDIA graphics card
- Not yet ranked on the public voice arenas
The details.
- Version
- Qwen3-TTS 1.7BReleased 2026-01
- Open source
- Yes, Apache-2.0
- Made by
- Alibaba (Qwen)
- Website
- github.com
Should you use it?
It is #9 of 11. Good for free voices on a gaming pc; for most people, ElevenLabs is the better start.
Can I run it myself?
Yes. Technical: command line or python.
Use qwen-tts Python package, mlx-audio (Mac).
- You need
- NVIDIA GPU with 8 GB, or an Apple Silicon Mac
- Setup
- TechnicalCommand line or Python
- Laptop, 8 GBToo small
- PC, 16 GB+Too small
- Mac, 16 GBRuns
- Mac, 32 GBRuns
- Mac, 64 GB+Runs
- PC, 8 GB GPURuns
- PC, 12 GB GPURuns
- PC, 16 GB GPURuns
- PC, 24 GB GPURuns
- WorkstationRuns
What it sounds like.
“Hi, I'm reading this sentence so you can compare AI voices side by side. Which one sounds the most human to you?”