text model · Mistral · macOS
Can I run Magistral Small on Apple M4 Pro (24GB)?
Yes. Magistral Small runs on Apple M4 Pro (24GB) at Q4_K_M (~16 GB of ~16 GB usable).
Fits at Q4_K_M (~16 GB of ~16 GB usable) but with little headroom. Close other apps, or drop to a 2k context to free about a gigabyte.
That figure is at a 4k context and moves about ±15% as context length changes.
- Q4_K_M needed
- ~16 GB
- Usable on device
- ~16 GB
- Device memory
- 24 GB
- Best quant
- Q4_K_M
Which quant fits
Running cost · estimate
- Power draw
- ~140 W
- Electricity / 1M tokens
- ~$0.36
- Pays for itself after
- ~14,279M tok
At ~$0.15/kWh and the estimated ~16 tok/s, a million generated tokens costs about $0.36 in electricity. Versus a hosted API at ~$0.5 per million tokens, the ~$1,999 Apple M4 Pro (24GB) pays for itself after roughly 14,279 million tokens, so local hardware is mostly a fixed cost, not a per-token one. TDP is the peak draw, so this is an upper bound. Assumptions.
Run it
Loads the Q4_K_M weights.
ollama run magistral:24b - Parameters
- 24B
- Q4_K_M size
- 14 GB
- Q8_0 size
- 25 GB
- Context
- 40k
- Ollama tag
- magistral:24b
- Memory
- 24 GB unified
- Usable for weights
- ~16 GB
- Power draw
- ~140 W
- Best runtime
- Ollama (MLX backend, preview) / MLX direct
You could also run
Run Magistral Small on other hardware
FAQ
Can Apple M4 Pro (24GB) run Magistral Small?
Yes. Magistral Small runs on Apple M4 Pro (24GB) at Q4_K_M (~16 GB of ~16 GB usable).
How much memory does Magistral Small need?
It is a tight fit on Apple M4 Pro (24GB). At Q4_K_M the weights are ~14 GB; with KV cache and runtime overhead, budget ~16 GB at a 4k context.
What is the best tool to run Magistral Small 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/magistral-small/apple-m4-pro-24gb) Sources
- apple.com/newsroom/2024/10/apple-introduces-m4-pro-and-m4-max
- apple.com/newsroom/2024/10/new-macbook-pro-features-m4-family-of-chips-and-apple-intelligence
- developer.apple.com
- github.com/ml-explore
- github.com/raullenchai
- huggingface.co
- lmstudio.ai
- ollama.com
- support.apple.com/en-us/103253
- support.apple.com/en-us/121553
- support.apple.com/en-us/121555
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.