localLLMriglocal, on your table

Hardware

Mac Studio M4 Max 128GB

70B Q4 and Q5 fit, Q8 often does too. Slower than a dual 3090, much quieter.

All hardware

Memory

128 GB unified

Bandwidth

546 GB/ssupport.apple.com/en-us/122211

Price

$3,683€3,290

14 Oct 2026

price unverified

Parts in the catalog

Which models fit

Estimate at 8,192 tokens, or the model's maximum when that is shorter. full means the formula fits in usable memory. offload means GPU plus system RAM. no means it does not fit. This is not a measured file.

ModelQ4estimateQ5estimateQ8estimateFP16estimate
Llama 3.2 3Bfullfullfullfull
Llama 3.1 8Bfullfullfullfull
Mistral 7B v0.3fullfullfullfull
Gemma 2 9Bfullfullfullfull
Qwen2.5 14Bfullfullfullfull
Phi-4 14Bfullfullfullfull
Gemma 2 27Bfullfullfullfull
Qwen2.5 32Bfullfullfullfull
Llama 3.3 70Bfullfullfullno
Qwen2.5 72Bfullfullfullno
DeepSeek-R1-Distill 70Bfullfullfullno

Open this in the calculator

Guides

Mac Studio M4 Max 128GB · localLLMrig