Mac Compute Rentals › Will it fit?
Will DeepSeek-V4-Pro-0813 fit on a 512GB Mac Studio (M3 Ultra)?
Not as published. Every MLX conversion we found is larger than the ~469 GiB GPU limit of a 512GB Mac.
deepseek-ai/DeepSeek-V4-Pro-0813: 1,650B parameters. License: mit.
| MLX weights | Size | On 512GB (~469 GiB GPU limit) |
|---|---|---|
| mlx-community/DeepSeek-V4-Pro-4bit | 779.5 GiB | Doesn't fit, 311 GiB over the limit |
| mlx-community/DeepSeek-V4-Pro-8bit | 791.3 GiB | Doesn't fit, 323 GiB over the limit |
| mlx-community/DeepSeek-V4-Pro | 791.3 GiB | Doesn't fit, 323 GiB over the limit |
Sizes: sum of .safetensors files from the Hugging Face API (Oct 2026). "GPU limit" is iogpu.wired_limit_mb=480000, already set on our Studios. Newer architectures may need a recent mlx-lm or llama.cpp; check before you book around one model.
Run it
uv venv ~/mlx && uv pip install --python ~/mlx/bin/python -U mlx-lm ~/mlx/bin/mlx_lm.chat --model <repo>
Try it on a whole 512GB M3 Ultra Mac Studio
3 days $299 · 7 days $489 · 14 days $779. Bare metal, one renter per machine, mlx-lm and llama.cpp ready.
Check availability & book