Hugging Face Inference Endpoints โ€” GPU rental, per GPU-hour

Cheapest on-demand rate per card, 8 cards.

GPUOn-demand / GPU-hour
NVIDIA T4
aws
$0.50
read
NVIDIA L4
aws
$0.80
read
NVIDIA L40S
aws
$1.80
read
NVIDIA A100 80GB
aws
$2.50
read
NVIDIA RTX PRO 6000 Blackwell
aws
$2.75
read
NVIDIA H100 80GB (form factor not stated)
aws
$4.50
read
NVIDIA H200
aws
$5.00
read
NVIDIA B200
aws
$9.25
read