MLX Community · First MLX build 13 Dec 2025
VibeVoice-Realtime-0.5B
MLX
Audio mit
A 1B parameter text to speech model, with 5 MLX builds, from 4-bit to fp16.
- Parameters
- 1B From Hugging Face
- Quantisation
- fp16 VibeVoice-Realtime-0.5B-fp16
- Memory needed
- 5 GB Estimate, unified memory
- Download
- 2.0 GB Weights, estimated
- Downloads
- 389 This build, last 30 days
- Task
- Text to speech Audio
- MLX builds
- 5 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 5 GB at fp16
8 GB
16 GB
24 GB
32 GB
64 GB
128 GB
Fits with headroomTight, close other appsWon’t fit
Estimated from the size of the weights, plus a fifth for context and working memory and 2 GB for macOS and your apps. Long contexts need more.
Run it
Terminal · Hugging Face CLI$ pip install -U huggingface_hub
$ hf download mlx-community/VibeVoice-Realtime-0.5B-fp16MLX builds
5 on mlx-community| VibeVoice-Realtime-0.5B-4bit | 4-bit | 633 MB | 3 GB | 141 |
|---|---|---|---|---|
| VibeVoice-Realtime-0.5B-5bit | 5-bit | 754 MB | 3 GB | 18 |
| VibeVoice-Realtime-0.5B-6bit | 6-bit | 876 MB | 3 GB | 20 |
| VibeVoice-Realtime-0.5B-8bit | 8-bit | 1.1 GB | 4 GB | 254 |
| VibeVoice-Realtime-0.5B-fp16 | fp16 | 2.0 GB | 5 GB | 389 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.