shisa-ai · First MLX build 11 Dec 2025
shisa-v2.1-llama3.3-70b
MLX
Text llama3.3
A 71B parameter text generation model from shisa-ai, with 2 MLX builds, from 8-bit to bf16.
- Parameters
- 71B From Hugging Face
- Quantisation
- 8-bit shisa-v2.1-llama3.3-70b-mlx-8bit
- Memory needed
- 86 GB Estimate, unified memory
- Download
- 75.0 GB Weights, estimated
- Downloads
- 100 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 86 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 · mlx-lm$ pip install -U mlx-lm
$ mlx_lm.generate --model mlx-community/shisa-v2.1-llama3.3-70b-mlx-8bit --prompt "Hello"MLX builds
2 on mlx-community| shisa-v2.1-llama3.3-70b-mlx-8bit | 8-bit | 75.0 GB | 86 GB | 100 |
|---|---|---|---|---|
| shisa-v2.1-llama3.3-70b-mlx-bf16 | bf16 | 141 GB | 160 GB | 77 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.