H100 80GB vs RTX 5080

For running local LLMs · 3 reports across 2 models

Side A

Vendor: nvidia
VRAM: 80GB
Memory: Discrete

Side B

Vendor: nvidia
VRAM: 16GB
Memory: Discrete

Tokens per second by model

Model	H100 80GB	RTX 5080
Qwen3.6up to 35B	45.0n=1	56.0n=1
Gemma 4up to 31B	125.3n=1	—

More comparisons