Aws Ec2 Cost Calculator
EC2 looks simple to price, an hourly rate times hours, until the pricing model enters the room. This AWS EC2 cost calculator prices one thing properly: instances. Pick a type and operating system, set a real usage schedule instead of assuming every box runs flat out, choose among On-Demand, Reserved, Spot and Savings Plan pricing, then add EBS storage and data transfer. The result shows effective hourly, daily, monthly and annual cost, prices all five models side by side on your exact configuration, and works out upfront payments and breakeven for Reserved terms.
The worked example used throughout is a single m5.large on Linux, On-Demand, running around the clock with 30 GB of gp3 storage: $71.75 per month. For the rest of an AWS bill, S3, RDS and Lambda, use the whole-estimate calculator linked further down; this page is EC2 depth, not breadth.
Planning estimate using us-east-1 rates. Reserved and Savings Plan commitments bill for the full term hours whether the instance runs or not; other regions run roughly 5 to 15% higher.
Quick Answer
Key takeaways
- Commitment models bill the full 730 hours a month whether the instance runs or not. This calculator prices them that way and warns you on screen when your schedule runs fewer hours than the commitment.
- Part-time schedules are where Reserved math breaks: an all-upfront 1-year Reserved box used 8 hours a day, 5 days a week needs 34.6 months to break even, longer than its own 12-month term.
- EBS storage bills while the instance is stopped, at $0.08 per GB-month for gp3. Stopping a box stops compute, never storage.
- Rates here are us-east-1 for every region selection, stated on the tool. Budget other regions about 5 to 15% higher.
- Quick Answer
- How to price an EC2 instance properly
- The EC2 cost formula
- Worked example: one m5.large at $71.75 a month
- The five pricing models, priced on the same box
- Why part-time instances change the commitment math
- EBS storage: the charge that survives a stopped instance
- Data transfer out of EC2, briefly
- What competing calculators miss
- Frequently asked questions
- Add This Calculator to Your Website
- References
How to price an EC2 instance properly
Set the instance first: type, count, operating system. Then be honest about the schedule, because hours are the lever most estimates fudge. A production box runs 24 x 7 x 4.3, about 722 hours a month. A dev box that sleeps nights and weekends runs closer to 172. Then pick the pricing model, add EBS storage per instance and expected transfer out, and press Calculate EC2 Cost. The comparison under the result reprices the same configuration on all five models, so the model decision and the instance decision stay separate, which is how AWS actually bills them.
The region picker needs one clarification, and the tool makes it on its own label: prices are us-east-1 rates for every selection. Treat the output as the baseline US East figure and scale up for other regions rather than mistaking the picker for regional pricing.
The EC2 cost formula
Three multiplications and one rule produce every figure in the result.
The hourly rates are AWS’s published us-east-1 figures, the same ladder the Amazon EC2 On-Demand pricing page publishes: an m5.large on Linux at $0.096 per hour, with Windows roughly doubling the small-instance rates because the license rides inside the hourly price.
Worked example: one m5.large at $71.75 a month
The calculator is pre-filled with the configuration most small production setups start from: one balanced instance, running continuously, modest SSD storage, no meaningful outbound transfer yet.
| Line | Math | Monthly |
|---|---|---|
| Compute (On-Demand) | $0.096 x 722.4 hrs (24 x 7 x 4.3) | $69.35 |
| EBS storage, 30 GB gp3 | 30 x $0.08 | $2.40 |
| Data transfer out | 0 GB: inside the 100 GB free allowance | $0.00 |
| Total | $2.36 per day, $861.00 per year | $71.75 |
Grow the same box and the lines move independently. At 100 GB of gp3 and 150 GB of monthly transfer out, storage becomes $8.00 and transfer becomes $4.50 (only 50 GB bills, after the free allowance), for a total of $81.85. Storage and transfer together are still a fraction of compute here, which stops being true for media and backup heavy workloads: at a terabyte of transfer a month, egress passes the instance itself.
The five pricing models, priced on the same box
Model choice is worth more than instance choice on steady workloads. Here is the worked example’s compute line under all five models, exactly as the calculator’s comparison prints it:
| Model | Factor | Compute / month | Saving vs On-Demand |
|---|---|---|---|
| On-Demand | 1.0, bills actual hours | $69.35 | baseline |
| Reserved 1-year | 0.68, bills 730 hrs | $47.65 | $21.70 (31%) |
| Reserved 3-year | 0.51, bills 730 hrs | $35.74 | $33.61 (48%) |
| Spot | 0.30, bills actual hours | $20.81 | $48.55 (70%) |
| Savings Plan | 0.75, bills 730 hrs | $52.56 | $16.79 (24%) |
Read the commitment rows carefully: Reserved and Savings Plan figures bill 730 hours even though the schedule runs 722.4, because the commitment does not pause. On-Demand suits spiky and short-lived work. Reserved rewards a known instance in a known region for one to three years. Savings Plans trade a little discount for freedom to change families and regions. Spot is the deepest discount in the table and the only one AWS can take back, with a two-minute interruption warning, so it belongs under batch jobs, queues and anything that checkpoints its own progress.
Commit (Reserved, Savings Plans)
- 31 to 48% off the worked example’s compute, every month, automatically.
- The bill runs at 730 hours whether the instance runs or not.
- Best after 2 to 3 months of stable usage data, not on day one.
Stay flexible (On-Demand, Spot)
- No term, no upfront, resize or terminate whenever the workload changes.
- On-Demand is full price; Spot is 70% off here but interruptible.
- Right for experiments, spiky traffic and fault-tolerant background work.
Why part-time instances change the commitment math
The most expensive EC2 mistake is a commitment bought for a box that sleeps. Put the worked example on an 8 hour, 5 day schedule and On-Demand compute falls to $16.51 a month, because it bills only the 172 hours used. A 1-year Reserved term for the same box still bills $47.65 a month at the full commitment, nearly three times the On-Demand cost of the actual usage. Pay that term all upfront, $571.85, and the breakeven against On-Demand stretches to 34.6 months on a term that only lasts 12. This calculator flags the situation in the result whenever a commitment model’s usage sits below 730 hours, and it is the check to run before any Reserved purchase: commitments reward hours you are certain of, and punish hours you merely hope for.
EBS storage: the charge that survives a stopped instance
Stopping an instance ends the compute meter and nothing else. Attached EBS volumes keep billing per GB-month at the volume rate, gp3 at $0.08, gp2 at $0.10, the io family at $0.125 plus $0.065 per provisioned IOPS, st1 at $0.045 and sc1 at $0.025. Snapshots taken from those volumes bill separately and accumulate quietly when retention is unmanaged. The practical rules are dull and effective: delete unattached volumes, set snapshot retention, and migrate any surviving gp2 volumes to gp3, which is cheaper per GB and faster at the baseline. When the question widens from one instance to the whole account, the aws cost calculator adds the S3, RDS and Lambda lines around this EC2 figure.
Data transfer out of EC2, briefly
Transfer into EC2 is free. Transfer out to the internet is free for the first 100 GB a month across the account, then $0.09 per GB at the first paid tier, which is the rate this calculator applies. The real schedule is tiered and falls to $0.05 per GB at very high volume, and routes like cross-AZ and NAT processing have their own charges that an instance-level estimate does not model. For transfer-heavy designs, the aws data transfer pricing calculator prices every route band by band. Pricing the equivalent virtual machine on the other major cloud is a reasonable final check: the azure vm pricing calculator covers that side.
What competing calculators miss
The official AWS Pricing Calculator prices EC2 correctly and buries the quick comparison most people came for under a multi-screen configuration flow. CostGoat is the closest rival here, with compute, storage and transfer on one screen, but it never explains the commitment-hours behavior that decides whether a Reserved price is real. Guide-led pages like Glassity and Wring explain the models well and calculate none of them. The systematic error in the quick tools is pricing commitments by usage hours: a part-time Reserved instance shown at its usage cost looks 60% cheaper than the commitment AWS will actually bill.
This page prices all five models on your schedule, bills commitments at the full 730 hours with an on-screen warning when you run less, and shows upfront breakeven in months, including when the honest answer is that the term never breaks even.
Frequently asked questions
On-Demand Linux in us-east-1, running around the clock, the compute is $69.35 per month at $0.096 per hour for 722.4 hours. Add storage and transfer for the real figure: the worked example with 30 GB of gp3 storage totals $71.75 per month, $861.00 per year. On commitment pricing the same compute is $47.65 on a 1-year Reserved term and $35.74 on a 3-year term.
Both commit you for one or three years and both bill for the full term whether the instance runs or not. A Reserved Instance locks a specific instance family, size and region for the deepest instance discount, and unused Reserved capacity can be resold on the RI Marketplace. A Savings Plan commits to a dollar-per-hour spend instead, and the discount follows your usage across families, sizes and regions, which makes it the more flexible default for fleets that change shape.
Yes. The commitment covers the full term hours, priced in this tool at 730 hours per month, regardless of whether a matching instance is running. That is why the calculator warns you when your usage schedule sits below the commitment: a 1-year Reserved m5.large used 8 hours a day, 5 days a week still bills $47.65 a month against $16.51 of On-Demand usage.
The first 100 GB per month is free, aggregated across the whole AWS account. Beyond that, this calculator applies $0.09 per GB, the first paid tier of the AWS schedule. The full schedule is tiered and drops to $0.085, $0.07 and $0.05 per GB at higher volumes, and the dedicated data transfer calculator on this site models every band and route.
Compute stops, storage does not. Attached EBS volumes keep billing per GB-month at their volume rate (gp3 $0.08) while the instance is stopped, snapshots keep billing too, and a public IPv4 address held by a stopped instance also bills by the hour. Only terminating the instance and deleting its volumes and snapshots ends the charges completely.
Spot instances run on spare AWS capacity at a steep discount, 70% off On-Demand in this tool’s worked comparison, and AWS can reclaim that capacity with a two-minute warning when the capacity is needed. Anything that cannot survive a two-minute shutdown does not belong on Spot. Batch processing, CI workers, rendering and queue-driven jobs that checkpoint progress are the natural fit.
By sticker rate, sc1 at $0.025 per GB-month, then st1 at $0.045, then gp3 at $0.08. In practice gp3 is the right default for almost every boot and application volume: it is solid state, includes a 3,000 IOPS baseline, and undercuts the older gp2 rate of $0.10. The io1 and io2 types at $0.125 plus $0.065 per provisioned IOPS are for databases that genuinely need guaranteed high IOPS.
All upfront converts the whole discounted term into one payment and sets the ongoing hourly charge to zero; partial upfront splits it half and half. Breakeven is the upfront amount divided by the monthly saving against On-Demand. On a full-time box it lands well inside the term. On the part-time example in the article, all upfront on a 1-year term breaks even at 34.6 months, beyond the term itself, which is the calculator’s way of saying do not buy that commitment.
Generally yes, typically about 5 to 15% more for the same instance, with some regions higher. This tool prices every region selection at us-east-1 rates and says so on the region field itself, so treat its output as the US East baseline. For a regional budget that has to be defended, confirm the figure in the official AWS Pricing Calculator with the region set.
Add This Calculator to Your Website
References
- Amazon Web Services, “Amazon EC2 On-Demand Pricing” – the hourly instance rates by type and operating system, and the data transfer schedule with its 100 GB free allowance. AWS: EC2 On-Demand Pricing. Accessed October 2026.
- Amazon Web Services, “Amazon EC2 Reserved Instances Pricing” – term, payment option and discount structure for Reserved Instances. AWS: Reserved Instances Pricing. Accessed October 2026.
- Amazon Web Services, “Amazon EC2 Spot Instances Pricing” – how Spot pricing works and what the discount buys. AWS: Spot Instances Pricing. Accessed October 2026.
- Amazon Web Services, “AWS Savings Plans” – the dollar-per-hour commitment model and its flexibility across families and regions. AWS: Savings Plans. Accessed October 2026.
- Amazon Web Services, “Amazon EBS Pricing” – per GB-month rates for gp3, gp2, io1, io2, st1 and sc1 volumes and provisioned IOPS charges. AWS: EBS Pricing. Accessed October 2026.
- Amazon Web Services, “AWS Pricing Calculator” – the official estimator for confirming instance pricing in a specific region before purchase. AWS Pricing Calculator. Accessed October 2026.