Everything on these pages is a public AWS list price, fetched by a script every night, run through the formulas below, and checked before it is published. This page is the audit trail.
| Number | Source | How |
|---|---|---|
| On-demand and reserved (RI) rates | AWS Price List API (pricing:GetProducts) | Linux / Windows / RHEL / SUSE, shared and dedicated tenancy, "Used" capacity, no pre-installed software; RDS by engine and deployment, ElastiCache by engine, OpenSearch and Redshift as sold. |
| Savings Plan rates | AWS Savings Plans offer file for the region (AWSComputeSavingsPlan) | The discounted hourly rate AWS publishes per instance SKU, for Compute and EC2 Instance Savings Plans, 1 and 3 years, no / partial / all upfront. |
| Spot | ec2:DescribeSpotPriceHistory | Seven days per availability zone; we publish the time-weighted median (p50), the minimum and the maximum. |
| Specs | ec2:DescribeInstanceTypes | vCPU, memory, processor, network, EBS, storage, GPU and the "current generation" flag exactly as EC2 reports them. Non-EC2 services use the Price List attributes. |
| Currency | European Central Bank | ECB euro foreign exchange reference rates of 2026-09-21, re-based to USD (1 USD = 95.820279 INR, 0.870322 EUR). Never a hard-coded rate. |
AWS quotes an RI as an upfront amount plus an hourly amount. We show one number per option: hourly + upfront ÷ (years × 8,760 h). That is the same convention the AWS Pricing Calculator uses for "effective hourly".
A month is 730 hours (8,760 ÷ 12). On-demand and spot are charged for the hours you run: rate × hours × quantity. A reserved instance or Savings Plan is charged for every hour of the term whether you use it or not: rate × 730 × quantity. That is why a commitment only wins above a certain number of hours per month: break-even hours = commitment rate × 730 ÷ on-demand rate. The calculator's "best option" is simply the option with the lowest monthly cost at the hours you entered, with spot shown separately because it can be interrupted.
24×7 = 730 h. Business hours = 10 h × 22 working days = 220 h. Nightly batch = 4 h × 30 nights = 120 h. Any other number can be typed in.
For a given type we list every type with the same family letter and the same size in a newer generation (m5.large → m6i.large, m7g.large …), its on-demand price, the price difference, and vCPU per dollar as a floor on the performance gain. A Graviton candidate is flagged when it means moving from x86 to arm64, because that needs arm64 builds.
The same type's on-demand rate in every region that sells it, with the monthly difference at 730 h. Instance price is only part of a bill: data transfer, latency and residency rules can outweigh it.
1,412 EC2 instance types across 34 regions at the last build. Prices are the property of Amazon Web Services; this calculator is not affiliated with AWS. Spotted a number that looks wrong? Tell us with the page link and we will check it against the source.
Tell us a little about your AWS setup and we will come back within one business day — or pick a slot straight away.