MLX Community · First MLX build 24 Apr 2026
DeepSeek-V4-Flash
MLX
Text
A 284B parameter text generation model, with 9 MLX builds, from 2-bit to bf16.
- Parameters
- 284B From Hugging Face
- Quantisation
- 2-bit DeepSeek-V4-Flash-2bit-DQ
- Memory needed
- 102 GB Estimate, unified memory
- Download
- 89.1 GB Weights, estimated
- Downloads
- 6.8K This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 9 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 102 GB at 2-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/DeepSeek-V4-Flash-2bit-DQ --prompt "Hello"MLX builds
9 on mlx-community| DeepSeek-V4-Flash-2bit-DQ | 2-bit | 89.1 GB | 102 GB | 6.8K |
|---|---|---|---|---|
| DeepSeek-V4-Flash-4bit | 4-bit | 160 GB | 181 GB | 4.9K |
| DeepSeek-V4-Flash-mxfp4 | MXFP4 | 160 GB | 181 GB | 97 |
| DeepSeek-V4-Flash-nvfp4 | 4-bit | 160 GB | 181 GB | 68 |
| DeepSeek-V4-Flash-5bit | 5-bit | 196 GB | 221 GB | 67 |
| DeepSeek-V4-Flash-6bit | 6-bit | 231 GB | 261 GB | 118 |
| DeepSeek-V4-Flash-8bit | 8-bit | 302 GB | 340 GB | 195 |
| DeepSeek-V4-Flash-mxfp8 | MXFP8 | 302 GB | 340 GB | 92 |
| DeepSeek-V4-Flash-bf16 | bf16 | 1123 GB | 1257 GB | 164 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.