How it works?

Many providers use a pay-as-you-use model and the cost entirely depends on your usage. The calculation is based on usage ($/invocation) and execution duration ($/hour). To find the compute costs, you first need to calculate monthly execution duration by multiplying number of executions with duration (s). Afterwards, multiply monthly duration with memory allocation (GB). You will get GB-s compute time per month which is used on all calculations.

Depending on the provider, every month you will get a number of invocations and seconds of compute time for free.

Table below shows the price for invocations and computational time.

Vendor 1m requests price GB/s of compute price
AWS Lambda $ $
Azure Functions $ (ceiled) $
Google Cloud Functions $ $
IBM Cloud Functions $ $