Compute Engine Cloud SQL BigQuery

Compute Engine costs

Resource Details

Estimation

Choose Your Commitment Plan

On-demand pricing
1yr Plan
- 37%
3yrs Plan
- 55%
Spot VMs
- 60-90%
ItemAmount
Add an instance to see an estimate.
Estimated Total
$0.00 / mo

Optimize Your GCP Spend

Connect your GCP project to Economize and discover savings opportunities up to 30% instantly.

Get started for free

HOW TO USE THIS CALCULATOR

Estimate your GCP bill in three steps

1

Configure your workloads

Select machine families (e.g. N1, N2, C2) or database configs and choose your GCP region. Rates vary by region — us-central1 is typically the cheapest baseline.

2

Set usage and quantity

Enter how many days per month each instance runs and how many instances you need. A 24/7 workload is 730 hours per month; dev environments often run far less.

3

Compare and decide

Evaluate Committed Use Discounts (1-yr or 3-yr CUDs) against On-Demand pricing to see instant savings and totals updated live.

PRICING MODEL

How Compute Engine pricing works

Compute Engine bills vCPU and memory as separate line items rather than charging a flat rate per instance, which is why two machine types with the same vCPU count can cost very different amounts. Predefined families such as E2, N2, N2D, C3, and the memory-optimized M series each carry their own per-vCPU and per-GB rates, and custom machine types are priced from the same components. Billing is per second after a one minute minimum, so short-lived batch jobs are charged only for what they use.

On top of the base rate, Google applies Sustained Use Discounts automatically once an instance runs for a significant share of the month, and Committed Use Discounts apply when you commit to one or three years of usage in a region. Spot VMs trade availability for a steep reduction and suit fault-tolerant work. The rates here cover the compute portion of the bill; persistent disks, snapshots, load balancing, and network egress are billed separately and should be added when you budget.

What drives your Compute Engine bill

Machine family and size

Each family targets a different workload profile. E2 is the cost-optimized general purpose option, N2 and N2D balance price and performance, C3 targets compute-bound work, and the M series is built for large in-memory databases.

Region

The same machine type varies noticeably between regions. us-central1 is usually among the cheapest, while Asia-Pacific and South American regions typically carry a premium for identical hardware.

Commitment term

Committed Use Discounts reduce the hourly rate in exchange for a one or three year commitment to a level of vCPU and memory in a region. Three year terms discount the most but lock capacity for longer.

Sustained use

Instances that run for a large portion of a month earn an automatic discount with no commitment and no configuration. This applies to eligible general purpose and memory-optimized families.

Spot capacity

Spot VMs use spare capacity at a large reduction and can be reclaimed at short notice. They fit batch processing, CI runners, and rendering, not stateful production services.

Compute Engine pricing questions

How accurate is this Compute Engine pricing calculator?
Rates are based on published Google Cloud on-demand and Committed Use Discount pricing for each region and are refreshed regularly. Your actual invoice may differ once persistent disks, snapshots, network egress, and support plans are included, since those are billed separately from compute.
How is the monthly Compute Engine cost calculated?
What are Committed Use Discounts and when do they pay off?
Are Sustained Use Discounts included in these estimates?
Why do Compute Engine prices differ between regions?
When should I use Spot VMs instead of on-demand instances?