OpenAI · First MLX build 14 Dec 2025
whisper-small
MLX
Audio apache-2.0
A 241M parameter speech recognition model from OpenAI, with 3 MLX builds, from 4-bit to fp16.
- Parameters
- 241M From Hugging Face
- Quantisation
- 4-bit whisper-small-4bit
- Memory needed
- 3 GB Estimate, unified memory
- Download
- 139 MB Weights, estimated
- Downloads
- 197 This build, last 30 days
- Task
- Speech recognition Audio
- MLX builds
- 3 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 3 GB at 4-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/whisper-small-4bitMLX builds
3 on mlx-community| whisper-small-4bit | 4-bit | 139 MB | 3 GB | 197 |
|---|---|---|---|---|
| whisper-small-8bit | 8-bit | 258 MB | 3 GB | 38 |
| whisper-small-fp16 | fp16 | 481 MB | 3 GB | 118 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.