Mistral AI · First MLX build 30 Jan 2025
Mistral-Small-24B-Instruct-2501
MLX
apache-2.0
A 24B parameter model from Mistral AI, with 5 MLX builds, from 3-bit to bf16.
- Parameters
- 24B From Hugging Face
- Quantisation
- 4-bit Mistral-Small-24B-Instruct-2501-4bit
- Memory needed
- 17 GB Estimate, unified memory
- Download
- 13.3 GB Weights, estimated
- Downloads
- 781 This build, last 30 days
- Task
- —
- MLX builds
- 5 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 17 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/Mistral-Small-24B-Instruct-2501-4bitMLX builds
5 on mlx-community| Mistral-Small-24B-Instruct-2501-3bit | 3-bit | 10.3 GB | 14 GB | 111 |
|---|---|---|---|---|
| Mistral-Small-24B-Instruct-2501-4bit | 4-bit | 13.3 GB | 17 GB | 781 |
| Mistral-Small-24B-Instruct-2501-6bit | 6-bit | 19.2 GB | 24 GB | 34 |
| Mistral-Small-24B-Instruct-2501-8bit | 8-bit | 25.0 GB | 30 GB | 82 |
| Mistral-Small-24B-Instruct-2501-bf16 | bf16 | 47.1 GB | 55 GB | 142 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.