DeepSeek · First MLX build 24 Apr 2026
DeepSeek-V4-Flash
MLX
Text mit
A 284B parameter text generation model from DeepSeek, with 8 MLX builds, from bf16 to 6-bit.
- Parameters
- 284B From Hugging Face
- Quantisation
- 8-bit DeepSeek-V4-Flash
- Memory needed
- 63 GB Estimate, unified memory
- Download
- 53.9 GB Weights, estimated
- Downloads
- 664 This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 8 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 63 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/DeepSeek-V4-Flash --prompt "Hello"MLX builds
8 on mlx-community| DeepSeek-V4-Flash-MTP-bf16 | bf16 | 26.4 GB | 32 GB | 130 |
|---|---|---|---|---|
| DeepSeek-V4-Flash | 8-bit | 53.9 GB | 63 GB | 664 |
| DeepSeek-V4-Flash-2bit-M-DQ | 2-bit | 89.1 GB | 102 GB | 133 |
| DeepSeek-V4-Flash-3bit-DQ | 3-bit | 125 GB | 142 GB | 104 |
| deepseek-ai-DeepSeek-V4-Flash-3bit | 3-bit | 125 GB | 142 GB | 101 |
| deepseek-ai-DeepSeek-V4-Flash-4bit | 4-bit | 160 GB | 181 GB | 243 |
| deepseek-ai-DeepSeek-V4-Flash-6bit | 6-bit | 231 GB | 261 GB | 133 |
| deepseek-ai-DeepSeek-V4-Flash-8bit | 8-bit | 302 GB | 340 GB | 630 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.