llamaperf

M5 Max 96GB vs RTX 3090

For running local LLMs · 7 reports across 2 models

Side A
M5 Max 96GB
Vendor
apple
VRAM
96GB
Memory
Unified
Side B
RTX 3090
Vendor
nvidia
VRAM
24GB
Memory
Discrete

Tokens per second by model

ModelM5 Max 96GBRTX 3090
Qwen3.6up to 35B72.9n=6
Qwen2.528.0n=1

More comparisons