H100 VRAM: 80 GB HBM3, and what fits in it

Hopper. 80 GB VRAM, verified from the vendor datasheet. On-demand rate from $3.00/hr, refreshed every 6 hours.

80 GB
VRAM
Hopper
Architecture
$3.00/hr
From
HBM3
Memory
Last updated: 2026-11-22 UTCRefreshes every 6 hours

Models that fit at FP16 (80 GB)

Models that fit at INT4 (quantized, 80 GB)

Caveat: Requires a quantized checkpoint actually published for the model, check its Hugging Face page before relying on this.

All models that fit in 80 GB: the open models whose weights and overhead fit, at native, FP8 and INT4 precision.

Live pricing

$3.00/hr on-demand rate, 7 regions with capacity, refreshed every 6 hours.

H100 guides

Start ComfyUI on a rented GPU.

© 2026 Aquanode. All rights reserved.

All trademarks, logos and brand names are the property of their respective owners.