llamaperf

RTX 3060 12GB vs RTX 5090

For running local LLMs · 15 reports across 2 models

Side A
RTX 3060 12GB
Vendor
nvidia
VRAM
12GB
Memory
Discrete
Side B
RTX 5090
Vendor
nvidia
VRAM
32GB
Memory
Discrete

Tokens per second by model

ModelRTX 3060 12GBRTX 5090
Qwen3.6up to 35B70.0n=33238.0n=7
Gemma 4up to 31B60.0n=2578.0n=3

More comparisons