Serverless Invocation Cost Calculator

Enter your monthly request volume and average function duration. See side-by-side costs on AWS Lambda, Cloudflare Workers, Vercel Functions, and Netlify — with free-tier allowances subtracted.

Workers CPU billing assumes 30% of wall-clock duration is billable CPU time. Lambda multiplies duration by allocated memory.

How serverless billing actually works

Serverless platforms split the bill into three buckets: a platform fee (some providers charge a monthly minimum once you leave hobby tiers), a request fee (per million invocations), and a compute fee (how long your code actually runs — metered differently on each platform).

On AWS Lambda, compute is GB-seconds: memory allocation times duration times invocation count. More RAM costs more even if your code does not use it. Cloudflare Workers bill CPU milliseconds instead — wall-clock time waiting on I/O is mostly free, which is why we model CPU at a fraction of your average duration.

Vercel and Netlify bundle invocations with GB-hours of active compute on their function plans. Cross the free-tier cliff and you need a paid plan before overages even matter — that step change is often bigger than the per-request delta.