Mistral AI · First MLX build 10 Dec 2025
Devstral-2-123B-Instruct-2512
MLX
Text other
A 125B parameter text generation model from Mistral AI, with 5 MLX builds, from 4-bit to bf16.
- Parameters
- 125B From Hugging Face
- Quantisation
- 4-bit Devstral-2-123B-Instruct-2512-4bit
- Memory needed
- 81 GB Estimate, unified memory
- Download
- 70.3 GB Weights, estimated
- Downloads
- 276 This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 5 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 81 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/Devstral-2-123B-Instruct-2512-4bit --prompt "Hello"MLX builds
5 on mlx-community| Devstral-2-123B-Instruct-2512-4bit | 4-bit | 70.3 GB | 81 GB | 276 |
|---|---|---|---|---|
| Devstral-2-123B-Instruct-2512-5bit | 5-bit | 86.0 GB | 99 GB | 123 |
| Devstral-2-123B-Instruct-2512-6bit | 6-bit | 102 GB | 116 GB | 107 |
| Devstral-2-123B-Instruct-2512-8bit | 8-bit | 133 GB | 151 GB | 141 |
| Devstral-2-123B-Instruct-2512-bf16 | bf16 | 250 GB | 282 GB | 91 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.