NVIDIA · First MLX build 8 Apr 2025
Llama-3_3-Nemotron-Super-49B-v1
MLX
Text nvidia-open-model-license
A 50B parameter text generation model from NVIDIA, with 2 MLX builds, from 4-bit to 6-bit.
- Parameters
- 50B From Hugging Face
- Quantisation
- 4-bit Llama-3_3-Nemotron-Super-49B-v1-mlx-4bit
- Memory needed
- 34 GB Estimate, unified memory
- Download
- 28.1 GB Weights, estimated
- Downloads
- 51 This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 2 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 34 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 · mlx-lm$ pip install -U mlx-lm
$ mlx_lm.generate --model mlx-community/Llama-3_3-Nemotron-Super-49B-v1-mlx-4bit --prompt "Hello"MLX builds
2 on mlx-community| Llama-3_3-Nemotron-Super-49B-v1-mlx-4bit | 4-bit | 28.1 GB | 34 GB | 51 |
|---|---|---|---|---|
| Llama-3_3-Nemotron-Super-49B-v1-mlx-6bit | 6-bit | 40.5 GB | 48 GB | 37 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.