Cloud Run Cost Calculator
Compare Cloud Run's requests-based and instance-based billing on your own traffic. Requests-based charges a third more per active second but nothing while idle; instance-based bills the whole lifetime and never charges per request.
Why, in one number
This service spends 100.0% of its instance time actually serving requests. That is high, so paying the requests-based premium on every active second costs more than instance-based billing wastes on idle ones. The two models tie at about 10% utilisation.
Requests-based
| CPU | $0.00 |
| Memory | $0.00 |
| Requests$0.40 per million | $1.20 |
| Idle instancesCPU throttled to a tenth of the active rate | $0.00 |
| Free tier | −$1.19 |
| Total | $1.20 |
Instance-based
| CPU | $0.00 |
| Memory | $0.00 |
| RequestsInstance-based billing has no per-request fee | Not charged |
| Idle instancesLifetime is billed uniformly | In the lines above |
| Free tier | −$0.30 |
| Total | $0.00 |
What this model assumes
- Rates are us-central1 (Iowa) (tier-1 pricing), list price, before any committed-use discount. GPU-attached services are not modelled.
- Concurrency divides the work, it does not multiply it. At concurrency 80, eighty simultaneous requests share one instance’s CPU allocation — so active vCPU-seconds are request-seconds divided by concurrency. Getting this backwards inflates an estimate by the concurrency factor, which is the most common error in Cloud Run maths.
- Requests-based billing charges CPU at $0.000024 per vCPU-second while a request is in flight and $0.0000025 when an instance is warm but idle. Instance-based charges $0.000018 for every second the instance exists, and nothing per request.
- The free tiers differ by model and both are applied to active usage first: instance-based grants 240,000 vCPU-seconds and 450,000 GiB-seconds; requests-based grants 180,000, 360,000, and two million requests.
- Startup time, cold starts, request-time CPU boost and outbound data transfer are not modelled. Egress in particular can exceed compute on a chatty service.
- Cloud Run CPU, memory and requests — source, read 4 Aug 2026
US East (N. Virginia) list prices, excluding any negotiated or committed-use discount. Your effective rate is lower if you have one.
Frequently asked questions
What is the difference between instance-based and requests-based billing on Cloud Run?
Requests-based billing charges CPU only while a request is being handled, at $0.000024 per vCPU-second, and adds $0.40 per million requests. Instance-based billing charges for every second the instance exists — busy or not — at $0.000018 per vCPU-second, and has no per-request fee. So requests-based costs a third more per active second but nothing for idle time, and instance-based is the reverse. Which is cheaper depends on what share of instance lifetime is spent actually serving.
Which Cloud Run billing model should I choose?
Roughly: a busy, steadily-loaded service is cheaper on instance-based, because it pays the lower rate on time it was going to use anyway and has little idle to waste. A spiky or low-traffic service is cheaper on requests-based, because instance-based would bill for a lot of time doing nothing. The crossover is a utilisation figure rather than a traffic level, and this calculator reports it for your own inputs.
How does concurrency affect Cloud Run cost?
Concurrency divides the work rather than multiplying it. At the default of 80, eighty simultaneous requests share one instance's CPU and memory allocation, so the vCPU-seconds you pay for are request-seconds divided by concurrency. Dropping concurrency to 1 — which some frameworks require — multiplies the compute bill by up to 80. It is the single largest lever on a Cloud Run bill and it is a one-line configuration change.
Do minimum instances cost money on Cloud Run?
Yes, and it is the most common surprise. A minimum instance is held warm to avoid cold starts, and it bills continuously. On requests-based billing its CPU is throttled and bills at $0.0000025 per vCPU-second — a tenth of the active rate — but its memory bills at the full rate regardless. On a low-traffic service, min-instances is frequently the largest line on the bill.
What does the Cloud Run free tier cover?
The allowances differ by billing model, which is itself a hint about the intended use of each. Instance-based grants 240,000 vCPU-seconds and 450,000 GiB-seconds a month. Requests-based grants 180,000 vCPU-seconds, 360,000 GiB-seconds and two million requests. Both are per billing account rather than per service, so a second service shares the same allowance.
One service is arithmetic. Two hundred is a job.
Finitizer reads every Cloud Run service in your organisation, measures the utilisation each one actually achieves, and flags the services on the wrong billing model, the ones holding warm instances they do not need, and the ones stuck at concurrency 1.