NVIDIA · First MLX build 4 May 2026
Nemotron-Mini-4B-Instruct
MLX
nvidia-open-model-license
A 4.2B parameter model from NVIDIA, with 5 MLX builds, from 4-bit to MXFP4.
- Parameters
- 4.2B From Hugging Face
- Quantisation
- 4-bit Nemotron-Mini-4B-Instruct-4bit-mlx
- Memory needed
- 5 GB Estimate, unified memory
- Download
- 2.4 GB Weights, estimated
- Downloads
- 186 This build, last 30 days
- Task
- —
- MLX builds
- 5 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 5 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/Nemotron-Mini-4B-Instruct-4bit-mlxMLX builds
5 on mlx-community| Nemotron-Mini-4B-Instruct-4bit-mlx | 4-bit | 2.4 GB | 5 GB | 186 |
|---|---|---|---|---|
| Nemotron-Mini-4B-Instruct-bf16-mlx | bf16 | 8.4 GB | 12 GB | 26 |
| Nemotron-Mini-4B-Instruct-nvfp4-4bit-mlx | 4-bit | 8.4 GB | 12 GB | 21 |
| Nemotron-Mini-4B-Instruct-mxfp4-4bit-mlx | MXFP4 | 8.4 GB | 12 GB | 15 |
| Nemotron-Mini-4B-Instruct-mixed-3-6-mlx | Unstated | 8.4 GB | 12 GB | 13 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.