Can an M3 Max 36GB Mac run Gemma 3 27B?

M3 Max · 36 GB unified memory · 300 GB/s · MacBook Pro 14" (M3 Max), MacBook Pro 16" (M3 Max)

Where the memory goes

Three things occupy memory when a model is loaded: the weights themselves, the KV cache that holds the conversation, and the runtime's own working buffers. Only the first is fixed — the KV cache grows with every token in your context.

The default. Best quality-per-gigabyte for most people.

How long a context fits

The KV cache is often what breaks a setup that looked fine at load time. Quantising it to 8-bit roughly halves its footprint with little measurable quality cost.

How to run it

Ollama is the shortest path. It picks a quantisation automatically — usually Q4_K_M, which is what the numbers above assume.

brew install ollama
ollama serve &
ollama run gemma3:27b

For MLX — Apple's own array framework, typically a little faster than llama.cpp on Apple Silicon:

pip install mlx-lm
mlx_lm.generate --model mlx-community/Gemma-3-27B-4bit --prompt "Hello"

Macs that run Gemma 3 27B comfortably

The smallest configuration that handles it is the M1 Max 32GB (MacBook Pro 14" (M1 Max)).

Common questions

Can an M3 Max 36GB Mac run Gemma 3 27B?

Yes. At Q4_K_M it needs about 21.1 GiB of the roughly 30.6 GiB available, generating around 14 tokens per second.

How fast is Gemma 3 27B on an Mac?

Around 14 tokens per second. Token generation on Apple Silicon is limited by memory bandwidth, and this machine has 300 GB/s.

How much memory does Gemma 3 27B need?

15.6 GiB for the weights at Q4_K_M, plus 3.9 GiB for an 8K-token KV cache and about 1.6 GiB of runtime overhead — 21.1 GiB in total.