llamaperf

M5 32GB vs RTX 4070

For running local LLMs · 2 reports across 1 model

Side A
M5 32GB
Vendor
apple
VRAM
32GB
Memory
Unified
Side B
RTX 4070
Vendor
nvidia
VRAM
12GB
Memory
Discrete

Tokens per second by model

ModelM5 32GBRTX 4070
Gemma 4up to 31Bn=155.0n=1

More comparisons