text model · Gemma · macOS
Can I run Gemma 4 31B on Apple M3 Pro (18GB)?
No. Gemma 4 31B needs ~20.5 GB even at Q4_K_M, but Apple M3 Pro (18GB) only has ~12 GB usable.
Needs ~20.5 GB even at Q4_K_M, but only ~12 GB is usable.
That figure is at a 4k context and moves about ±15% as context length changes.
The gap is about 8.5 GB: Gemma 4 31B needs roughly 20.5 GB at Q4_K_M and Apple M3 Pro (18GB) leaves only about 12 GB usable for a model. The lightest tracked hardware that runs Gemma 4 31B is the Nvidia GeForce RTX 4090 (24GB) at 24 GB. See Gemma 4 31B on Nvidia GeForce RTX 4090 (24GB).
- Q4_K_M needed
- ~20.5 GB
- Usable on device
- ~12 GB
- Device memory
- 18 GB
Which quant fits
- Parameters
- 32.7B
- Q4_K_M size
- 18.32 GB
- Q8_0 size
- 32.64 GB
- Context
- 256k
- Ollama tag
- gemma4:31b
- Memory
- 18 GB unified
- Usable for weights
- ~12 GB
- Power draw
- ~70 W
- Best runtime
- Ollama (llama.cpp Metal backend) / MLX
What you can run instead
Run Gemma 4 31B on other hardware
FAQ
Can Apple M3 Pro (18GB) run Gemma 4 31B?
No. Gemma 4 31B needs ~20.5 GB even at Q4_K_M, but Apple M3 Pro (18GB) only has ~12 GB usable.
How much memory does Gemma 4 31B need?
Apple M3 Pro (18GB) does not have enough memory. At Q4_K_M the weights are ~18.32 GB; with KV cache and runtime overhead, budget ~20.5 GB at a 4k context.
What is the best tool to run Gemma 4 31B 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/gemma-4-31b/apple-m3-18gb) 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.