M5 32GB vs RTX 3090

For running local LLMs · 12 reports across 3 models

Side A

Vendor: apple
VRAM: 32GB
Memory: Unified

Side B

Vendor: nvidia
VRAM: 24GB
Memory: Discrete

Tokens per second by model

Model	M5 32GB	RTX 3090
Qwen3.6up to 35B	—	104.0n=9
Qwen2.5	—	70.0n=2
Gemma 4up to 31B	—n=1	—

More comparisons