Builds / Apple M4 (unified)

Mac mini M4 (16 GB): Qwen3.5 9B Q4_K_M at 17.7 tok/s, Gemma 4 12B at 13.1 (llama.cpp Metal)

Last updated: 1 Nov 2026

Source: mustafa.net ↗📏 How measured: reported by the original source; RigBase re-checked the numbers at that source

GPU(s): Apple M4 (unified)
VRAM / unified memory (GB): 16
CPU: Apple M4
Motherboard: not stated
RAM: 16 GB unified memory
PSU: not stated
Other: Official llama.cpp build for macOS, run directly on the Apple GPU
Reported total cost: not stated
Power: not stated

Measured speed (generation tokens/sec)

Model	Quant	Engine	tok/s	Prompt processing (tok/s)	Details
Qwen3.5 9B	Q4_K_M	llama.cpp server (Metal)	17.7	not stated	context: 8,192 ctx, short fixed prompt📏 llama.cpp server, 8,192 ctx, 256 tokens at temperature 0; middle of three runs after a warm-up (runs agreed within ~1%)
Gemma 4 12B	Q4_K_M	llama.cpp server (Metal)	13.1	not stated	context: 8,192 ctx, short fixed prompt📏 llama.cpp server, 8,192 ctx, 256 tokens at temperature 0; middle of three runs after a warm-up (runs agreed within ~1%)
Qwen3.5 4B legacy	Q4_K_M	llama.cpp server (Metal)	29.7	not stated	context: 8,192 ctx, short fixed prompt📏 llama.cpp server, 8,192 ctx, 256 tokens at temperature 0; middle of three runs after a warm-up (runs agreed within ~1%)

Mustafa.net homelab test, measured 12 Oct 2026 (published 13 Oct 2026). Same model files, prompt and settings as the RTX 2060 Super build from the same article. The author left Mac memory use out as unreliable. The same article also ran CPU-only boxes (Core i7-8559U NUC: 3.1 tok/s, Xeon E5-2665 with 4-channel DDR3: 4.2 tok/s on Qwen3.5 9B).

Source: https://mustafa.net/2026/10/13/local-llm-calculator-tested-four-machines/

Blank fields = not stated in the source. We never estimate build data; GPU street prices above are separately sourced and dated.

Embed this result

Copy into a README or blog. Links back to this build. See badge docs (CC BY 4.0 attribution). The SVG does not invent speeds — numbers stay on the build page.

Markdown

[![Measured on RigBase](https://getrigbase.com/badge/mac-mini-m4-16gb-qwen35-gemma4-mustafa.svg)](https://getrigbase.com/builds/mac-mini-m4-16gb-qwen35-gemma4-mustafa/)

HTML

<a href="https://getrigbase.com/builds/mac-mini-m4-16gb-qwen35-gemma4-mustafa/" rel="noopener"><img src="https://getrigbase.com/badge/mac-mini-m4-16gb-qwen35-gemma4-mustafa.svg" alt="Measured on RigBase" height="20"/></a>

Preview: · How to use

Rent before you buy: try this model on a cloud GPU for a few dollars first: Runpod (referral link; sign up with Google to get the new-user bonus)

Quick answers: Can Apple M4 run Gemma 4 12B? · Can Apple M4 run Qwen3.5 9B?