Cloud testing tools charge in four main ways: per usage unit (pay-as-you-go), by subscription tier, by committed spend, and by running open source on your own cloud account. For load testing the usage unit is usually the virtual user hour (VUH); for device and browser clouds it is a parallel session or a seat. This article explains each model, shows the arithmetic with clearly labelled illustrative numbers, and prints a few list prices that were read from vendor pricing pages on 23 October 2026.
Key takeaways
- Compare models by the unit being metered (virtual user hours, tests, parallel sessions, seats), not by the headline monthly price.
- Pay-as-you-go suits irregular testing; a subscription suits regular testing. The break-even is a simple division, shown below.
- Free tiers and trials are the cheapest way to learn how many units your own tests really use.
- Vendor prices change often. The figures here are dated; confirm on the vendor’s pricing page before deciding.
What pricing models do cloud testing tools use?
| Model | You pay for | Typical unit | Main risk |
|---|---|---|---|
| Pay-as-you-go | What you consume | Virtual user hour, test, minute | A bill that grows with every extra run |
| Subscription tiers | A bundle of limits each month or year | Concurrent users, tests per month, test duration, seats | Paying for limits you do not use, or hitting a cap mid-month |
| Committed use | A yearly minimum in exchange for a lower rate | Annual commitment, often custom quoted | The commitment outlasts your needs |
| Open source on your own cloud | Compute you rent from a cloud provider, plus your own time | Instance hours | Engineering time to run and scale it |
Most vendors mix these. A common shape is a monthly plan with an included allowance and a metered rate above it.
How does pay-as-you-go pricing work?
You are billed for what you use. In load testing the usual unit is the virtual user hour: virtual users multiplied by the time they run. In device and browser clouds it is typically a session, a minute or a parallel slot.
Illustrative arithmetic (invented numbers, not a vendor’s rates): a test with 200 virtual users that runs for 30 minutes uses 200 x 0.5 = 100 VUH. If that run cost $0.20 per VUH, it costs $20. Ten such runs in a month cost $200. A quiet month with two runs costs $40. That swing is the point of the model: it is cheap when you test rarely and uncertain when a release week triggers many reruns.
Check three things before relying on it: whether ramp-up and ramp-down time count as billable time, whether the rate drops at higher volume, and whether the vendor caps spend.
How do subscription tiers work?
A subscription gives you a fixed monthly or annual price for a bundle of limits. Load testing plans are usually limited by concurrent virtual users, tests per month, maximum test duration and parallel tests, with higher tiers adding VUH allowances, data retention, team features and support levels.
LoadFocus pricing follows this pattern: paid plans for load testing are limited mainly by concurrent virtual users, tests per month and per year, test duration and parallel tests, billed monthly or yearly with no minimum contract term and a 7-day free trial on paid plans. This article prints no LoadFocus prices; the product pages are the source for those.
The upside is a budget you can forecast. The downside is the cap: if your next test needs more virtual users than the tier allows, you move up a tier, not up a little.
When does a subscription beat pay-as-you-go?
Divide the flat fee by the metered rate to find the break-even usage. Using invented numbers again: a plan at $100 a month against a metered rate of $0.20 per VUH breaks even at 100 / 0.20 = 500 VUH a month. Below 500 VUH, pay-as-you-go is cheaper; above it, the plan is. If the plan also includes an allowance, subtract it first.
Real plans rarely compare this cleanly because they meter different things (tests, concurrent users, VUH). Convert your own typical test into each vendor’s unit first, then compare.
What do real load testing prices look like? (read 23 October 2026)
These are list prices from each vendor’s own pricing page on 23 October 2026, in US dollars as shown. They exclude taxes, discounts and add-ons, and will change.
| Vendor and plan | Model | Figure read from the page |
|---|---|---|
| Grafana Cloud k6, Free | Included allowance | 500 virtual user hours a month |
| Grafana Cloud k6, Pro | Platform fee plus metered | From $0.15 per VUH plus a $19 monthly platform fee that covers the first 500 VUH; the page says the rate drops with volume |
| Grafana Cloud k6, Enterprise | Committed use | Custom pricing with a $25,000 a year minimum commitment |
| BlazeMeter Performance, Basic | Subscription tier | $149 monthly or $1,188 billed annually; 1,000 concurrent users, 15 tests a month, 1 hour maximum test |
| BlazeMeter Performance, Pro | Subscription tier | $649 monthly or $5,988 billed annually; 5,000 concurrent users, 6,500 VUH a month, 5 hours maximum test |
| BrowserStack Automate, Chrome Desktop | Per parallel session, annual billing | $59 a month as listed for annual billing (device cloud, not load testing) |
Worked example on Grafana’s published numbers: 1,000 VUH in a month would be the $19 fee covering the first 500, plus 500 x $0.15 = $75, so about $94 a month at the starting rate. The page says the rate falls with volume, so treat that as an upper bound. The same 1,000 VUH on a different vendor’s tier model cannot be priced until you know its concurrency and test-count limits.
The table also shows how different the units are: VUH on one, tests and concurrent users on another, parallel sessions on a third. That is why headline comparisons mislead. For the wider open source against commercial question, including what self-hosting costs in engineering time, see the open source vs commercial load testing tools comparison, which this article deliberately does not repeat.
Why do load tests cost more than other cloud tests?
Because the vendor has to run the load generators. A functional test on a browser grid occupies one browser. A load test with thousands of virtual users needs many generator machines, often in several regions, for the whole run. Pricing therefore tracks concurrency and duration. The related question of where the generators run is covered in cloud load testing vs on-premise.
How do you estimate your own bill?
- Write down your typical test: virtual users, duration, runs per week.
- Convert it to units: VUH = virtual users x hours; also note the maximum concurrent users you will ever need.
- Count reruns. Release weeks usually need several runs of the same test.
- Price it in each model with the break-even division above.
- Use a free plan or trial to measure real usage before committing. Cloud load testing with LoadFocus starts with a free plan, and the 7-day trial on paid plans is stated on its pricing page.
- Review quarterly. Usage changes with team size and release cadence.
Which model should you choose?
- Pay-as-you-go if tests are rare, project-based or still being explored.
- A subscription if tests run on every release or in CI, and the limits (users, duration, parallel tests) fit what you run.
- A committed contract only when volume is high and predictable enough that a discount outweighs the lock-in; vendors quote these individually.
- Open source if you have people to run it. See the comparison of cloud testing tools for how the options differ.
Frequently Asked Questions
What is the difference between pay-as-you-go and subscription pricing for cloud testing?
Pay-as-you-go bills for the units you consume, such as virtual user hours or test minutes. A subscription bills a fixed fee for a bundle of limits. The first suits irregular testing and the second suits regular testing, and the break-even is the flat fee divided by the metered rate.
What is a virtual user hour?
A virtual user hour is one simulated user running for one hour, so 200 virtual users for 30 minutes is 100 VUH. Vendors differ in whether ramp periods count, so read the metering rules.
Are free plans enough to evaluate a load testing tool?
They are enough to measure how many units your test uses and to check that scripts run. Free plans usually cap concurrency, duration or monthly allowance, so they rarely cover a full-scale rehearsal.
How do overages work?
Some plans stop at the cap and ask you to upgrade; others bill the excess at a metered rate. LoadFocus states that usage beyond included API and browser monitoring runs is billed as metered overage; check each vendor’s terms for load tests.
Can I mix pay-as-you-go and a subscription?
Yes. A base plan for routine tests plus metered usage for launch weeks is a common shape, and Grafana Cloud k6’s Pro plan is built that way.
Sources
Read on 23 October 2026 and named as plain text: Grafana pricing page (Free, Pro, Enterprise); BlazeMeter pricing page (Performance plans); BrowserStack pricing page (Automate); LoadFocus pricing page (structure, billing, trial). Arithmetic examples marked “invented” or “illustrative” are not vendor figures.
Published with PostNext
