llamaperf

M3 Ultra 256GB vs M4 Max 128GB

For running local LLMs · 7 reports across 5 models

Side A
M3 Ultra 256GB
Vendor
apple
VRAM
256GB
Memory
Unified
Side B
M4 Max 128GB
Vendor
apple
VRAM
128GB
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 Ultra 256GBM4 Max 128GB
GLM-5.3up to 321B37.4n=1n=1
Qwen3.8up to 2400B15.2n=2
Minimax M317.3n=1
Museup to 30B25.0n=1
Qwen3.6up to 35B72.5n=1

More comparisons