text model · DeepSeek-R1 · macOS
Can I run DeepSeek-R1-0528 on Apple M4 (24GB)?
No. DeepSeek-R1-0528 needs ~384.1 GB even at Q4_K_M, but Apple M4 (24GB) only has ~16 GB usable.
Needs ~384.1 GB even at Q4_K_M, but only ~16 GB is usable.
That figure is at a 4k context and moves about ±15% as context length changes.
The gap is about 368.1 GB: DeepSeek-R1-0528 needs roughly 384.1 GB at Q4_K_M and Apple M4 (24GB) leaves only about 16 GB usable for a model. No single tracked device has enough memory; DeepSeek-R1-0528 needs a multi-GPU or high-memory rig.
- Q4_K_M needed
- ~384.1 GB
- Usable on device
- ~16 GB
- Device memory
- 24 GB
Which quant fits
- Parameters
- 671B (MoE, 37B active)
- Q4_K_M size
- 377.13 GB
- Q8_0 size
- 664.3 GB
- Context
- 160k
- Ollama tag
- deepseek-r1:671b
- Memory
- 24 GB unified
- Usable for weights
- ~16 GB
- Power draw
- ~65 W
- Best runtime
- Ollama (MLX backend, preview) / MLX direct
What you can run instead
FAQ
Can Apple M4 (24GB) run DeepSeek-R1-0528?
No. DeepSeek-R1-0528 needs ~384.1 GB even at Q4_K_M, but Apple M4 (24GB) only has ~16 GB usable.
How much memory does DeepSeek-R1-0528 need?
Apple M4 (24GB) does not have enough memory. At Q4_K_M the weights are ~377.13 GB; with KV cache and runtime overhead, budget ~384.1 GB at a 4k context. It is a Mixture-of-Experts model (671B total / 37B active), so all experts must stay in memory; memory tracks total params, not active params.
What is the best tool to run DeepSeek-R1-0528 on macOS?
LM Studio for a simple setup; mlx-lm for the most speed. vLLM is NOT a Mac tool, it is a CUDA/Linux serving engine. Unified memory is not a fixed VRAM slice; ~70% is usable for weights.
Embed this
A live badge for your model card or README, updated as the data is. [](https://localmodel.run/can-i-run/deepseek-r1-0528/apple-m4-24gb) Sources
- aider.chat
- apple.com
- developer.apple.com
- github.com/ml-explore
- github.com/raullenchai
- huggingface.co/deepseek-ai/DeepSeek-R1-0528
- huggingface.co/deepseek-ai/DeepSeek-R1-0528/blob
- huggingface.co/unsloth/DeepSeek-R1-0528-GGUF/tree/main/BF16
- huggingface.co/unsloth/DeepSeek-R1-0528-GGUF/tree/main/Q4_K_M
- huggingface.co/unsloth/DeepSeek-R1-0528-GGUF/tree/main/Q8_0
- lmstudio.ai
- ollama.com
- openrouter.ai
- stencel.io
- support.apple.com/en-us/103253
- support.apple.com/en-us/121555
- support.apple.com/en-us/122209
Weights are measured from GGUF files; KV cache and overhead are computed, so totals can vary ~15% with context and runtime. Any tok/s is a bandwidth estimate. See methodology.