stepfun-ai · First MLX build 7 Oct 2026
Step-Audio-EditX
MLX
Audio apache-2.0
A 3.5B parameter audio to audio model from stepfun-ai, with 2 MLX builds, from 8-bit to bf16.
- Parameters
- 3.5B From Hugging Face
- Quantisation
- 8-bit Step-Audio-EditX-8bit
- Memory needed
- 7 GB Estimate, unified memory
- Download
- 3.8 GB Weights, estimated
- Downloads
- 55 This build, last 30 days
- Task
- Audio to audio Audio
- MLX builds
- 2 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 7 GB at 8-bit
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/Step-Audio-EditX-8bitMLX builds
2 on mlx-community| Step-Audio-EditX-8bit | 8-bit | 3.8 GB | 7 GB | 55 |
|---|---|---|---|---|
| Step-Audio-EditX-bf16 | bf16 | 7.1 GB | 10 GB | 50 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.