llamaperf

M3 Max 128GB vs M3 Ultra 192GB

For running local LLMs · 3 reports across 2 models

Side A
M3 Max 128GB
Vendor
apple
VRAM
128GB
Memory
Unified
Side B
M3 Ultra 192GB
Vendor
apple
VRAM
192GB
Memory
Unified

Highest reported generation speed by model family

These maxima come from different community setups. Model size, quantization, GPU count, offloading and concurrency may differ. They do not establish which GPU is faster for the same workload. Check the comparison conditions →

ModelM3 Max 128GBM3 Ultra 192GB
Qwen3.6up to 35B5.5n=2
DeepSeek V4 Flashup to 285B43.0n=1

More comparisons