H100 80GB vs RTX 3090

For running local LLMs · 7 reports across 2 models

Side A

Vendor: nvidia
VRAM: 80GB
Memory: Discrete

Side B

Vendor: nvidia
VRAM: 24GB
Memory: Discrete

Tokens per second by model

Model	H100 80GB	RTX 3090
Qwen3.6up to 35B	45.0n=1	66.0n=5
Qwen2.5	—	28.0n=1

More comparisons