Source review date: 10 November 2026
Reference rate transparency
The four-provider engine includes example hourly compute rates, storage unit rates, request charges, free allowances and basic network rates. Some example values were checked against official documentation on the reference date; others are clearly labelled illustrative inputs and should never be treated as provider quotations. Rates change by contract, region, resource size, architecture and time.
| Provider | Compute reference shape | Verified reference rate fields | Documentation |
|---|---|---|---|
| Amazon Web Services | t3.medium (2 vCPU, 4 GiB) | compute, egress, freeEgress | Official provider source ↗ |
| Microsoft Azure | B2s-like planning baseline | None — editable illustrative only | Official provider source ↗ |
| Google Cloud | e2-medium-like planning baseline | object, egress, freeEgress, functions, freeRequests | Official provider source ↗ |
| Oracle Cloud Infrastructure | E4.Flex 1 OCPU + 4GB | compute, freeEgress | Official provider source ↗ |
How the estimate works
For each provider, modeled recurring monthly spending equals reference VM count × billed VM hours × editable VM hourly rate, plus managed DB quantity × billed DB hours × editable DB hourly rate, plus object and block storage GB × their rates. Egress charges apply to modeled outbound GB after the configured allowance, while serverless requests apply beyond the specified free request count.
A support reserve can be applied as a percentage of the direct line-item subtotal. Optional contingency and tax are then calculated in sequence. Your growth forecast increases storage, cross-region transfers, internet egress and serverless requests monthly, while holding the number of always-on VMs and managed database instances constant.
Scope exclusions
The model does not fully price: spot instances; commitments and reservations; Windows and commercial licensing; request-heavy storage operations; database I/O and replicas; snapshots and backup retention; Kubernetes orchestration; dedicated load balancers; NAT gateways; content delivery networks; observability; security tools; cross-availability-zone distinctions; third-party support; function duration, CPU and RAM; progressive transfer billing bands; or all country-specific taxes.
Why the four reference shapes differ
AWS T3 is burstable, the Azure B2s planning profile is burstable, Google E2-medium is a shared-core reference, and OCI E4.Flex uses OCPUs. Equal vCPU labels do not mean equal compute performance, baseline rights or sustained capacity. The comparison shows how selected example assumptions affect spending, not a guaranteed performance-normalized comparison.
Official provider pricing links
- AWS Pricing
- Microsoft Azure Pricing
- Google Cloud Pricing Calculator
- Oracle Cloud Price List
- Cloudflare D1 Pricing
- OpenAI API Pricing
- Anthropic Claude Pricing
- Gemini API Pricing
Before publishing a cost recommendation to a client, validate exact regional pricing directly with the vendor, confirm the expected workload, identify which free quotas apply at account level, and state the date and currency of the estimate.