What is RunPod?
RunPod is an all-in-one cloud built for AI, providing a globally distributed GPU cloud for developing, training, and scaling AI models. It offers a seamless experience for machine learning workflows, allowing users to focus on building models rather than infrastructure.
Features of RunPod
Develop
RunPod provides a managed and community-driven template repository, allowing users to deploy any container on the AI cloud. Public and private image repositories are supported, and users can configure their environment as needed.
Train
RunPod offers a range of NVIDIA H100s and A100s, as well as AMD MI300Xs and AMD MI250s, for machine learning training tasks. Users can reserve GPUs up to a year in advance.
Autoscale
RunPod's serverless GPU workers scale from 0 to n with 8+ regions distributed globally. Users only pay when their endpoint receives and processes a request.
Cost-Effective
RunPod provides a cost-effective platform for developing and scaling machine learning models, with zero fees for ingress/egress and 99.99% guaranteed uptime.
How to Use RunPod
Spin up a Pod
Users can spin up a GPU pod in seconds, with a choice of 50+ templates ready out-of-the-box or the option to bring their own custom container.
Deploy
RunPod allows users to deploy any container on the AI cloud, with public and private image repositories supported. Users can configure their environment as needed.
Scale
RunPod's serverless GPU workers scale from 0 to n with 8+ regions distributed globally, ensuring that users can handle large volumes of requests.
Pricing
RunPod offers a range of pricing options, including:
Secure Cloud
Starting from $2.89/hour for an H100 PCIe with 200 GB disk, 1 x H100 PCIe, 9 vCPU, and 50 GB RAM.
Community Cloud
Starting from $1.19/hour for an A100 PCIe with 80 GB VRAM, 83 GB RAM, and 8 vCPUs.
Serverless
Pricing is based on the number of requests processed, with a minimum charge of $0.05/GB/month for network storage.
Helpful Tips
Zero Ops Overhead
RunPod handles all operational aspects of infrastructure, from deploying to scaling, allowing users to focus on building models.
Easy-to-use CLI
RunPod's CLI tool allows users to automatically hot reload local changes while developing and deploy on Serverless when done.
Secure & Compliant
RunPod AI Cloud is built on enterprise-grade GPUs with world-class compliance and security.
Frequently Asked Questions
What is the cold-start time for RunPod's serverless workers?
RunPod's serverless workers have a cold-start time of sub 250 milliseconds, thanks to Flashboot.
Is RunPod secure and compliant?
Yes, RunPod AI Cloud is built on enterprise-grade GPUs with world-class compliance and security. It is in the process of getting SOC 2, ISO 27001, and HIPAA certifications.