Run Gemma 4 31B on M4 Max

Yes — Gemma 4 31B (31B) runs at 18 tok/s on M4 Max with 48 GB RAM using Q4_K_M quantization via MLX. First token latency is 1.1s. Google's flagship 31B dense model, Arena AI #3 among open models.

Speed
18
tok/s
First Token
1.1
seconds
RAM Needed
48
GB minimum
Engine
MLX
recommended
Want to run Gemma 4 31B faster — or step up to bigger models? Find your Mac →

Benchmark Details

The LLMCheck index estimates Gemma 4 31B on M4 Max using our published methodology: Q4_K_M quantization, memory-bandwidth scaling, and cross-referenced third-party benchmarks where available. Figures are transparent estimates — own this config? Submit a real benchmark →

MetricValue
Tokens per second18 tok/s
Time to first token1.1s
QuantizationQ4_K_M
Minimum RAM48 GB
Recommended engineMLX
Parameters31B
Benchmark date2026-04

Q4_K_M 31B MLX M4 Max

Setup Guide: Run Gemma 4 31B on M4 Max

The recommended engine for Gemma 4 31B on M4 Max is MLX. Install with pip and pull the model:

pip install mlx-lm
mlx_lm.generate --model mlx-community/gemma-4-31b-q4_k_m --prompt "Hello!"

Alternatively, you can use Ollama for a simpler setup:

ollama run gemma4:31b

Performance on Other Apple Silicon Chips

ChipSpeedFirst TokenMin RAMEngine
M5 Max 26 tok/s 0.7s 128 GB MLX
M5 Max 22 tok/s 0.9s 64 GB Ollama
M4 Pro 14 tok/s 1.4s 24 GB Ollama

System Requirements

To run Gemma 4 31B on M4 Max you need:

🛒 Get a Mac that runs Gemma 4 31B

Gemma 4 31B needs about 48 GB of unified memory. These current Apple Silicon Macs have the headroom to run it comfortably:

As an Amazon Associate, LLMCheck earns from qualifying purchases. These affiliate links cost you nothing extra and help keep our benchmarks free.

Compare More Models

See how Gemma 4 31B stacks up against other models on your specific Mac hardware.

Open Compare Tool Full Leaderboard