text model · Mistral · macOS
Can I run Devstral Small on Apple M4 (24GB)?
Yes. Devstral Small runs on Apple M4 (24GB) at Q4_K_M (~15.4 GB of ~16 GB usable).
Fits at Q4_K_M (~15.4 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. Apple M4 (24GB) leaves ~0.6 GB of headroom.
- Q4_K_M needed
- ~15.4 GB
- Usable on device
- ~16 GB
- Device memory
- 24 GB
- Best quant
- Q4_K_M
Which quant fits
Running cost · estimate
- Power draw
- ~65 W
- Electricity / 1M tokens
- ~$0.39
- Pays for itself after
- ~11,809M tok
At ~$0.15/kWh and the estimated ~7 tok/s, a million generated tokens costs about $0.39 in electricity. Versus a hosted API at ~$0.5 per million tokens, the ~$1,299 Apple M4 (24GB) pays for itself after roughly 11,809 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
Pick your tool. All 3 load the same Q4_K_M weights.
ollama run devstral:24b llama-cli -hf mistralai/Devstral-Small-2507_gguf:Q4_K_M lms get mistralai/Devstral-Small-2507_gguf - Parameters
- 24B
- Q4_K_M size
- 13.35 GB
- Q8_0 size
- 23.33 GB
- Context
- 128k
- Ollama tag
- devstral:24b
- Memory
- 24 GB unified
- Usable for weights
- ~16 GB
- Power draw
- ~65 W
- Best runtime
- Ollama (MLX backend, preview) / MLX direct
You could also run
Run Devstral Small on other hardware
FAQ
Can Apple M4 (24GB) run Devstral Small?
Yes. Devstral Small runs on Apple M4 (24GB) at Q4_K_M (~15.4 GB of ~16 GB usable).
How much memory does Devstral Small need?
It is a tight fit on Apple M4 (24GB). At Q4_K_M the weights are ~13.35 GB; with KV cache and runtime overhead, budget ~15.4 GB at a 4k context.
What is the best tool to run Devstral 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/devstral-small/apple-m4-24gb) Sources
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.