Amazon AWS

What's good about Amazon AWS

• Broad range of compute options including GPUs
• Highly scalable and reliable infrastructure
• Pay-as-you-go pricing with cost optimization tools
• Extensive global network of data centers
• Rich ecosystem of integrated services and tools

Amazon AWS GPU prices

10 GPU models across 19 configurations. Rank compares Amazon AWS's cheapest on-demand listing for each GPU with every provider's cheapest.

What a run costs on Amazon AWS

24 hours at today's cheapest listing for each GPU, at the size it is listed.

• 1× H100 SXM for 24 hours

  us-east-1 at $6.88/GPU-hr

  $165.12

  $83.76 over the market median rate

• 8× A100 SXM 80GB for 24 hours

  us-east-1 at $3.43/GPU-hr

  $658.73

  $349.61 over the market median rate

• 8× H200 SXM for 24 hours

  us-east-1 at $7.91/GPU-hr

  $1,519.10

  $654.46 over the market median rate

The market median is every provider's cheapest on-demand price for that GPU.

Amazon AWS LLM API prices

Pay-per-token pricing, per 1M tokens.

Prices last updated November 13, 2026

Billing & terms

On-Demand Instances

Pay for compute capacity by the second with no long-term commitments.

Spot Instances

Use spare EC2 capacity at up to 90% off the On-Demand price.

Reserved Instances

Save up to 72% compared to On-Demand pricing with a 1 or 3-year commitment.

Savings Plans

Save up to 72% on compute usage with a 1 or 3-year commitment to a consistent amount of usage.

EC2 Capacity Blocks for ML

Reserve accelerated compute capacity for a future start date and a defined duration; billed as an upfront reservation fee plus an operating system fee.

Regions

39 geographic regions and 123 availability zones worldwide.

Support

Basic (free), Developer, Business, Enterprise support plans with varying response times and features. Extensive documentation, forums, and training resources.

Watch out for

• Complex pricing structure
• Steep learning curve for new users
• Potential for unexpected costs without proper management

Features

Global Infrastructure

Extensive network of data centers across multiple regions worldwide

Pay-as-you-go Pricing

Flexible pricing model with no upfront commitments required

Advanced Security

Comprehensive security tools and compliance certifications

Auto Scaling

Automatically adjust resources based on demand

Integrated Services

Extensive ecosystem of services that work seamlessly together

Developer Tools

Comprehensive suite of tools for development, deployment, and management

AWS Free Tier

Sign-up credits and no-cost usage limits for evaluating services before committing spend

Services

• Amazon EC2

  Virtual servers in the cloud with a wide range of instance types.

• Amazon ECS

  Fully managed container orchestration service.

  • Support for Docker containers
  • Integration with other AWS services
  • Automated cluster management and scheduling

• Amazon EKS

  Managed Kubernetes service for container orchestration.

  • Certified Kubernetes conformant
  • Integrates with AWS networking and security services
  • Supports both EC2 and Fargate launch types

• AWS Lambda

  Serverless compute service for running code without managing servers.

  • Automatic scaling and high availability
  • Pay only for compute time consumed
  • Supports multiple programming languages

• Amazon Bedrock

  Fully managed service providing access to foundation models from AI21 Labs, Anthropic, Cohere, Meta, Mistral, Stability AI, and Amazon through a unified API.

  • Multi-Provider Access: Access models from multiple AI providers through a single API
  • Model Customization: Fine-tune models on your own data with continued pretraining
  • Guardrails: Built-in safety controls and content filtering
  • AgentCore: Platform for running production AI agents built with any framework or model, with managed authentication, access control, and tracing
  • On-Demand: Pay per input/output token with no minimum commitment
  • Provisioned Throughput: Reserved model units for consistent performance

Getting started

1. Step 1

   Sign up for AWS

   Create an AWS account to access the cloud platform.

2. Step 2

   Choose a compute service

   Select from EC2, Lambda, or container services based on your workload needs.

3. Step 3

   Launch an instance

   Configure and launch your first compute instance or container.

4. Step 4

   Set up security

   Configure security groups and access controls for your resources.

5. Step 5

   Monitor and optimize

   Use AWS CloudWatch and Compute Optimizer to monitor performance and reduce costs.

Tools that work with Amazon AWS

Coding agents, gateways and other software that can use Amazon AWS, listed by name. All tools

• AiderAI coding agents
• AnythingLLMAI chat apps
• Claude CodeAI coding agents
• ClineAI coding agents
• Cloudflare AI GatewayLLM gateways and routers
• dstackGPU orchestration
• HeliconeLLM gateways and routers
• Kilo CodeAI coding agents
• LangfuseLLM observability and evals
• LangSmithLLM observability and evals
• LibreChatAI chat apps
• LiteLLMLLM gateways and routers
• MLflowExperiment tracking
• OpenCodeAI coding agents
• PortkeyLLM gateways and routers
• SkyPilotGPU orchestration