In this serverless compute model, you get billed for the time your functions are running and depending on the resources you allocated to your functions. You can find more info at https://cloud.google.com/functions/pricing

The nearest 100ms means if your function runs in 230ms you get billed for 300ms.