4-bit Open on Hugging Face
Parameters
672B
Weights (est.)
≈ 378 GB
Mac memory (est.)
512 GB+
Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.
Run it
With mlx-lm on an Apple silicon Mac:
pip install mlx-lmmlx_lm.generate --model mlx-community/DeepSeek-V3.2-4bit --prompt "Hello"
Details
- Base model
- deepseek-ai/DeepSeek-V3.2 (finetune)
- Vendor
- DeepSeek
- Task
- Text generation
- Quantisation
- 4-bit
- Licence
- mit
- Downloads
- 609 in the last 30 days · 19K all time · 7 likes
- Published
- 2025-12-01 · updated 2026-01-06 (10mo ago)
- Tags
- deepseek_v32
Other MLX builds of deepseek-ai/DeepSeek-V3.2
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| DeepSeek-V3.2-4bit (this one) | 4-bit | ≈ 378 GB | 512 GB+ | 609 |
| DeepSeek-V3.2-mlx-5bit | 5-bit | ≈ 462 GB | > 512 GB | 100 |
| DeepSeek-V3.2-8bit | 8-bit | ≈ 714 GB | > 512 GB | 169 |
| DeepSeek-V3.2_bf16 | bf16 | ≈ 1344 GB | > 512 GB | 130 |