Runpod

What it does

Runpod is a cloud platform for running AI workloads on rented GPUs and CPUs. You bring containers, frameworks, and code, then run them as dedicated Pods, autoscaling Serverless endpoints, or multi-node Clusters. Public Endpoints add API access to pre-deployed image, video, audio, and language models, authenticated with an API key.

Pods suit development and long-running jobs, Serverless suits production inference that scales to zero when idle, and Clusters suit distributed training with high-speed networking between nodes. In Serverless, you write a handler and deploy it, and Runpod handles queuing and distribution across workers. Storage includes container disk, volume disk, and network storage.

Runpod suits developers and ML teams that want to rent GPU capacity by the second or hour rather than buy hardware. Using it requires a Runpod account and console access, and API calls require an API key. Pricing varies by GPU type, deployment model, and region, and reserved or enterprise capacity requires contacting sales. Confirm current rates and regional availability before committing a workload.

What it can do

• Launch GPU Pods with preloaded environments across 30+ GPU models and 31 global regions
• Run autoscaling Serverless GPU endpoints that scale to zero when idle
• Deploy multi-node Clusters for distributed training and large-batch inference
• Call pre-deployed image, video, audio, and language models through Public Endpoints
• Attach container, volume, and network storage to workloads
• Use managed logs and metrics for deployed endpoints

Example workflows

Serve an image model as a Serverless endpoint

Input: a ComfyUI or model handler container. Steps: write the handler, push it to Serverless, configure scaling. Output: a live inference endpoint that autoscales and bills per use.

Call a pre-deployed model via Public Endpoints

Input: an API key and a text prompt with width and height. Steps: send a POST request to the model's runsync URL with the prompt in the input object. Output: a generated image URL in the response.

Before you choose

• Pricing depends on GPU type, deployment model, storage, and region; check the pricing page for current rates.
• Reserved and enterprise capacity require contacting sales; the self-serve pricing page covers on-demand use.
• Compliance details (SOC 2 Type II, HIPAA, GDPR) are documented in the Trust Center and vary by region and deployment model.

Listed under

• Coding
• Sandboxes and infrastructure
• AI infrastructure
• Developers
• Fine-tuning
• GPU cloud
• GPU pods
• Model training

Get started

Create an account and deploy a Pod or Serverless endpoint.

1. Create a Runpod account and sign in to the console.
2. Create an API key from the console if calling Public Endpoints or Serverless via API.
3. Choose a GPU Pod, Serverless endpoint, or Cluster based on workload duration and traffic pattern.

Pods bill per hour or per second by GPU type; Serverless bills inference workers per usage; Clusters billed per GPU-hour with reserved options via sales. Storage is billed per GB per month. Confirm exact rates on the pricing page.

Source: runpod.io/pricing, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• The AI Developer Cloud | Runpod ↗www.runpod.io
• GPU Cloud Pricing | Per-Second H100, A100, RTX | Runpod ↗www.runpod.io
• Welcome to Runpod - Runpod Documentation ↗docs.runpod.io

Company website

runpod.io: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.