Model · OpenBMB · Text-To-Speech
VoxCPM2
OpenBMB's open text-to-speech model — 30 languages, voice cloning from a short clip, and Voice Design that invents a new voice from a written description. Apache 2.0, so a business can ship it without a licence fee.
- Modality
- Text-To-Speech
- License
- Apache 2.0 (Open)
- Parameter size
- 2B
- Released
- April 6, 2026
- Last verified
- October 28, 2026
- Runs locally
- Yes
What you need to run it
- Mac
- Yes, but slow. Any Apple Silicon Mac (M1 or newer) with at least 16 GB (24 GB recommended) of memory. It runs, but much slower than on an NVIDIA card.
- Windows / Linux PC
- NVIDIA works best. A graphics card with at least 8 GB of video memory.
- No graphics card
- Not practical. Too slow without a graphics card or Apple Silicon.
Works on a Mac through PyTorch MPS or a Metal GGUF build, but at about 1.8x slower than real time on an M4 Pro, so an NVIDIA card is much faster.
Numbers are for: official 2B weights. Based on the maker’s own figures; where the maker gives only a graphics-card figure, the Mac memory is our estimate.
Strengths
- Apache 2.0 for both weights and code — free for commercial use
- 30 languages with no language tag needed in the input text
- Voice Design: describe a voice in words (age, tone, pace) and get it, with no reference audio
- Clones a voice from a short reference clip, with optional style guidance for emotion and pace
- 48 kHz output and real-time streaming (real-time factor about 0.3 on an RTX 4090)
- Runs on a graphics card with about 8 GB of memory
Weaknesses
- Needs an NVIDIA GPU with CUDA 12 or newer for the reference setup
- Voice Design and controllable cloning vary between runs — OpenBMB suggests generating one to three times
- Realistic cloning carries impersonation risk; the README forbids fraud and impersonation use
Try it
| Where | Type | Notes |
|---|---|---|
| Hugging Face | weights | Apache 2.0 |
| GitHub (voxcpm package) | local | Python package and CLI; also served by vLLM-Omni and a llama.cpp fork |
Official sources
- Model card model
- GitHub github
- Documentation docs
- Technical report paper
Change log
- — Added hardware requirements (Mac support, memory, NVIDIA). Source: https://github.com/OpenBMB/VoxCPM.
- — Initial entry. Added 2026-10-28 from the popularity sweep: 40.8k GitHub stars, 983k Hugging Face downloads/month, 1,732 likes. The README dates VoxCPM2 only to April 2026; the release date used is the 2.0.0 GitHub release tag (2026-04-06). Replaces VoxCPM-0.5B (2025-09) and VoxCPM1.5 (2025-12), which supported only Chinese and English.
Esc