RTX 3090 vs RTX 4080

For running local LLMs · 3 reports across 2 models

Side A

Vendor: nvidia
VRAM: 24GB
Memory: Discrete

Side B

Vendor: nvidia
VRAM: 16GB
Memory: Discrete

Tokens per second by model

Model	RTX 3090	RTX 4080
Qwen3.6up to 35B	66.0n=2	—
Qwen2.5	28.0n=1	—

More comparisons