# Overview - Documentation

> For the complete documentation index, see [llms.txt](/llms.txt). Markdown is available with `Accept: text/markdown` and `.md` URL variants.

Source: /docs/guides/gpu-instance-overview

# Overview

GPU Instance provides cost-effective, accessible high-performance computing with easy GPU access, making it ideal for complex AI tasks. Users benefit from on-demand rental, optimizing both efficiency and cost-effectiveness.

##

[​](#features)

Features

- Rich built-in templates help you quickly deploy GPU instances for AI tasks out of the box;

- Start a GPU instance in seconds, helping you quickly handle traffic peaks;

- Provides nodes in multiple global regions to deploy GPUs closer to your users for minimal latency;

- Provides free local storage and allows you to attach network volume, also provides free network transfer;

- Billing is accurate to the second, charging only for the actual running time of the GPU container instances;

- Affordable pricing save costs by up to 50%.

##

[​](#terminology)

Terminology

- Container Image: Docker-compatible OCI image reference, supporting both public registries and private repositories with authentication.

- Environment Variables: Runtime configuration key-value pairs injected into containers for dynamic configuration and secrets management.

- HTTP/TCP Ports: Network endpoints exposed by containers for communication. HTTP ports (e.g., 80, 443) for web services, TCP ports for general network protocols.

- Container Disk: Root filesystem partition containing OS and system binaries. Mounted as read-write at container startup.

- Volume Disk: High-performance ephemeral storage directly attached to the GPU instance. Optimized for I/O intensive workloads but non-persistent across instance restarts.

- Network Volume: Network-attached storage service that allows users to access and manage data through a network. Data can be stored on remote servers and accessed from any device.

If you have any questions, you can first check the [FAQs for GPU Instance](/docs/guides/faq#gpu-instance).

Last modified on August 6, 2025