GPU directory

Cloud GPU prices,
compared.

9 GPU models from 25 active cloud providers. Live $/hr pricing, ranked cheapest first.

01 The market right now

VRAM × Price

Find the sweet spot.

Each dot is one GPU. Higher VRAM to the right, higher $/hr toward the top. Coral = datacenter, lavender = workstation, sage = consumer, slate = mobile.

All tiers Datacenter Workstation Consumer · 12–24 GB 24–48 GB 48–80 GB 80GB+ Clear all

GPU	Tier	Architecture	VRAM	Providers	AI models	From
Nvidia RTX A2000	workstation	Ampere	12GB	2 →	6 →	$0.022/hr	Compare →
Nvidia RTX A4500	workstation	Ampere	20GB	2 →	6 →	$0.025/hr	Compare →
Nvidia RTX A4000	workstation	Ampere	16GB	4 →	36 →	$0.026/hr	Compare →
Nvidia T4	datacenter	Turing	16GB	4 →	11 →	$0.070/hr	Compare →
Nvidia RTX 4000 Ada SFF	workstation	Ada Lovelace	20GB	1 →	112 →	$0.17/hr	Compare →
Nvidia RTX 4000 Ada	workstation	Ada Lovelace	20GB	1 →	107 →	$0.19/hr	Compare →
Nvidia Tesla V100	datacenter	Volta	16GB	9 →	6 →	$0.21/hr	Compare →
Nvidia Nvidia RTX A4000	workstation	Ampere	16GB	1 →	6 →	$0.32/hr	Compare →
Nvidia RTX 2000 Ada	workstation	Ada Lovelace	16GB	1 →	6 →	$0.46/hr	Compare →

Cloud GPU prices,compared.

Find the sweet spot.

Cloud GPU prices,
compared.