AWS EC2 Pricing and Billing: The Meters Around the Instance-Hour
Ten m5.large instances for 730 hours is $700.80. The 60-second floor and the meters around the instance-hour are what the rate omits.
EC2 charges for instance runtime, not CPU utilization. The example prices ten m5.large instances at an editable $0.096 per hour for 730 hours: $700.80 before discounts.
Every result is pre-rendered with the default inputs, so the numbers are readable with JavaScript off. Results update as you type and nothing leaves your browser. Part of the Buildopsy engineering calculators.
Price a fleet from instance count, hourly rate and the fraction of the month it is running, then add short-launch billing and attached EBS storage. Apply an optional effective discount scenario. The rates, the 60-second floor and the meters this model leaves out are documented in the EC2 pricing and billing teardown.
Two meters sit outside the instance price. Per-second billing carries a 60-second minimum, so a launch shorter than a minute still bills a full minute. Attached EBS storage bills by provisioned GB-month whether the instance runs or is stopped. Both default to zero, so the total above does not change until you edit them.
A launch below 60 seconds is billed as 60. Ten thousand 30-second launches run 83.3 hours but bill 166.7, doubling the hours on that line. SUSE Linux Enterprise Server is the exception: AWS bills SLES instances as a full hour, so the 60-second floor does not apply. The $0.08 EBS rate is the gp3 example AWS publishes for its pricing walkthrough; check your Region.
What this estimate includes, and what it leaves out
| Meter | Modeled here |
|---|---|
| Instance-hours, on-demand | Yes — running share of the month and short launches at the 60-second minimum |
| Attached EBS volume storage | Yes — provisioned GB-month at the rate you enter |
| Data transfer out | No — priced separately by the egress calculator |
| EBS snapshots, provisioned IOPS and throughput | No |
| Elastic IPs, load balancers, NAT gateways | No |
| Reserved Instance and Savings Plan billing | No — the discount field is a scenario, not a quote |
Per-second billing with a 60-second minimum is stated on the AWS On-Demand pricing page. EBS per-second billing and the gp3 example rate come from the Amazon EBS pricing page, re-checked October 28, 2026.
Ten m5.large instances for 730 hours is $700.80. The 60-second floor and the meters around the instance-hour are what the rate omits.
Multiply instance count by the hourly rate and billable running hours. This modeled example uses $0.096 per hour for 730 hours, or $70.08 per instance before any user-entered discount. Attached EBS storage and short-lived launches are added separately when you enter them.
No. EC2 instance usage is billed while the instance is running, even if it is idle. The running-time input models powered-on time as a share of the entered hours, not CPU load.
AWS bills On-Demand usage by the second with a 60-second minimum, so any launch shorter than a minute is charged for the full minute. The calculator applies that floor, which is why ten thousand 30-second launches bill 166.7 instance-hours instead of 83.3. SUSE Linux Enterprise Server is the exception: AWS bills SLES instances as a full hour.
The instance charge stops when it is stopped, but attached EBS volumes keep billing by provisioned GB-month until you delete them, and any Elastic IP or snapshot charges continue. The EBS field models the storage part of that.
No. The discount field applies an effective percentage to modeled running usage. Reserved Instances and Savings Plans are commitments that can be billed even when matching usage is absent, and actual discounts depend on terms and usage.
AWS S3 Storage Cost Calculator · Cloud Egress Cost Calculator · Kubernetes Cluster Cost Calculator · Kubernetes HPA Replica Calculator · Downtime Cost Calculator