Which GPU should I use for AI?

We have compiled some AI GPU benchmarks for training popular models on our AI GPU benchmarks page.

Can I run a GPU for an hour to test it?

Once you have created a billing account and added credit, you can test a GPU instance for as long as you like.

Do you offer any discounts?

We offer discounts for commited use terms. Learn more about our commitment pricing on our pricing page.

GPU cloud

Easily rent cloud GPUs for a variety of workloads

Select from range of high performance and cost effective GPUs that fit all your needs. Perfect for AI and workloads.

Launch an instance Speak with an expert

Build a GPU cloud that’s perfect for your workloads

Deploying AI based workloads on CUDO Compute is easy and cost-effective. Follow our AI related tutorials.

Tutorial

Deploying LLM’s like Google Gemma

In this tutorial we will run Google Gemma with Ollama so that you can send queries via a Rest API

Read tutorial

Tutorial

Deploy PyTorch

With CUDO Compute you can deploy PyTorch docker containers to the latest NVIDIA Ampere Architecture GPUs.

Read tutorial

Instantly deploy a range of high-performance GPUs

Access a wide range of performant NVIDIA and AMD GPUs to accelerate your AI, ML & HPC workloads

GPU model	On-demand cost	Committed cost
H100 SXM80GB HBM2e (3.35 TB/s bandwidth)	from $2.45/hr	from $1.80/hr Save $26,040.96	More info Deploy
H100 PCIe94GB HBM2e (3.9 TB/s bandwidth)	from $2.45/hr	from $2.15/hr Save $13,147.20	More info Deploy
A100 PCIe80GB HBM2e (1.9 TB/s bandwidth)	from $1.50/hr	from $1.25/hr Save $10,956.00	More info Deploy
L40S48GB GDDR6 (864 GB/s bandwidth)	from $0.88/hr	from $0.75/hr Save $3,419.52	More info Deploy
A800 PCIe80GB HBM2e (1.94 TB/s bandwidth)	from $0.80/hr	from $0.76/hr Save $87.36	More info Deploy
RTX A600048GB GDDR6 (768 GB/s bandwidth)	from $0.45/hr	from $0.40/hr Save $109.20	More info Deploy
A4048GB GDDR6 (696 GB/s bandwidth)	from $0.39/hr	from $0.35/hr Save $87.36	More info Deploy
V10016GB HBM2 (900 GB/s bandwidth)	from $0.39/hr	from $0.23/hr Save $4,103.42	More info Deploy
RTX A500024GB GDDR6 (768 GB/s bandwidth)	from $0.35/hr	from $0.30/hr Save $109.20	More info Deploy
H200 SXM141GB HBM3e (4.8 TB/s bandwidth)	POA	-	More info Deploy
RTX 4000 SFF Ada20GB GDDR6 (280 GB/s bandwidth)	POA	-	More info Deploy

NVIDIA H100 SXM

from $2.45/hr

Deploy performant H100s on-demand with CUDO Compute.

Learn more Deploy

NVIDIA H100 PCIe

from $2.45/hr

Deploy performant H100s on-demand with CUDO Compute.

Learn more Deploy

NVIDIA A100 PCIe

from $1.50/hr

Deploy performant A100s on-demand with CUDO Compute.

Learn more Deploy

NVIDIA L40S

from $0.88/hr

Deploy performant L40Ss on-demand with CUDO Compute.

Learn more Deploy

NVIDIA A800 PCIe

from $0.80/hr

Deploy performant A800s on-demand with CUDO Compute.

Learn more Deploy

NVIDIA RTX A6000

from $0.45/hr

Deploy performant RTX A6000s on-demand with CUDO Compute.

Learn more Deploy

NVIDIA A40

from $0.39/hr

Deploy performant A40s on-demand with CUDO Compute.

Learn more Deploy

NVIDIA V100

from $0.39/hr

Deploy performant V100s on-demand with CUDO Compute.

Learn more Deploy

NVIDIA RTX A5000

from $0.35/hr

Deploy performant RTX A5000s on-demand with CUDO Compute.

Learn more Deploy

NVIDIA H200 SXM

Pricing on request.

Deploy performant H200s on-demand with CUDO Compute.

Learn more Deploy

NVIDIA RTX 4000 SFF Ada

Pricing on request.

Deploy performant RTX 4000 SFF Adas on-demand with CUDO Compute.

Learn more Deploy

Virtual machines

Deploy GPU workloads on virtualised infrastructure

Deploy cost-effective virtual machines on our platform for a wide range of workloads.

Learn more

Bare metal

Control your infrastructure with GPUs on bare metal

Give yourself more control over your infrastructure with our on-demand bare metal

Learn more

Effortlessly manage your GPU infrastructure

Automatically provision GPU resources, monitor performance and optimise costs.

Realtime monitoring

Gain real-time insights into your GPU usage and resource allocation, allowing you to optimise your AI workflows. Track key performance indicators, identify bottlenecks and proactively manage your resources for maximum efficiency.

Global network

Access a geographically distributed network of high-performance GPUs in multiple regions. Leverage our extensive global infrastructure to model training and inference anywhere in the world.

Cost management

Benefit from transparent pricing models and detailed billing reports. Easily track your cloud resource expenditure and identify opportunities for cost optimisation.

Commitment pricing

Save up to 30% on GPU costs by committing to a fixed term. Our commitment pricing plans offer significant savings, ideal for long-term projects and workloads.

Reserve GPUs, deploy clusters and scale with our enterprise solutions

View enterprise solutions Speak with an expert

Frequently asked questions

Which GPU should I use for AI?: We have compiled some AI GPU benchmarks for training popular models on our AI GPU benchmarks page.
Can I run a GPU for an hour to test it?: Once you have created a billing account and added credit, you can test a GPU instance for as long as you like.
Do you offer any discounts?: We offer discounts for commited use terms. Learn more about our commitment pricing on our pricing page.

Are you looking for support with something more specific? Check out our knowledge base