TestFinch

BlogFlowQA

AI testing tool pricing in 2026: what 15 tools charge for AI test generation and test runs

AI testing tool pricing comes in six units: per step, per test run, credits, per seat, per maintained test and a flat managed fee. One table of what 15 tools publish on their own pricing pages, checked on 18 November 2026, and one small suite priced on each.

AI testing tool pricing in 2026 runs from $0 to $2,500 a month on published plans, and the tools charge in six different units, so two prices rarely compare directly. AI test generation is priced three ways: included in a plan (FlowQA, QA.tech, Checksum), drawn from credits or tokens (Stably, TestSprite, KaneAI, TestGrid, Functionize, Qase, Kualitee), or charged per AI-generated step (Momentic, at $0.025). Eight of the 15 tools in the table have a free plan, and three publish no price for a paid one.

TestFinch makes FlowQA. Every other figure comes from the vendor's own page, linked in the table and checked on 18 November 2026.

What 15 AI testing tools charge#

Tool How it charges Free plan Lowest published paid price What that buys Source, checked 18 Nov 2026
FlowQA Per cloud test run; runs on your own computer are free Yes $49 a month 3,000 cloud test runs and 3 explorations a month, up to 5 members Pricing
Momentic Per step, plus browser minutes Yes, $50 of usage a month $500 a month $500 of usage at $0.0125 a step, or $0.025 for an AI-generated or failure-recovery step momentic.ai/pricing
Stably Credits for browser minutes and AI tokens Yes, $10 of credits a month $60 a month $60 of credits, at $0.01 a browser minute and $0.30 to $15 per million tokens stably.ai/pricing
TestSprite Credits, on plans billed per seat Yes, 150 credits a month $19 a month, first month $0 400 credits a month; the page does not say what one credit buys testsprite.com/pricing
Functionize Studio Credits for agent work Yes, 800,000 credits a month $20 a month 3,200,000 credits a month; extra credits are $10 per 200,000 functionize.com/pricing
KaneAI by TestMu AI Per agent, with credits No $19 per agent a month, or about $17 billed annually 2,000 credits a month and tests written and run on your own browser testmuai.com/pricing
QA.tech Credits, used only by test runs Trial with 10 credits $89 a month 50 credits a month, up to 5 users and 2 applications; the page does not say what one credit buys qa.tech/pricing
QA Wolf Per test run Free runs for new teams; the number is not stated $1 per web run One run of up to 5 minutes, attempted up to three times qawolf.com/pricing
TestGrid Per seat, with tokens for AI test creation No; a guided proof of concept $249 per seat a month 4 devices or browsers and 5,000 tokens; a typical test uses 10 to 20 tokens to create testgrid.io/pricing
Bug0 Flat managed fee No; a discounted 60-day pilot $2,500 a month An engineer and unlimited runs, up to 500 user flows, month to month bug0.com/pricing
mabl Credits for cloud runs; local runs are free 14-day trial Not published The page names 500 credits a month as a starting point and offers a quote mabl.com/pricing
Checksum Per maintained workflow 30-day trial Not published Tiers of 50, 200 and 400 or more workflows, each a test of up to 30 actions checksum.ai/product/pricing
testRigor Pay as you go, rate not shown Yes, with every test and result public Not published Enterprise is custom pricing testrigor.com/sign-up
Qase Per user, with AI credits Yes, without AI credits $35 per user a month billed yearly, or $42 paid monthly Test management with 2,000 AI credits a month; extra credits are $0.40 each qase.io/pricing
Kualitee Per user, with AI credits Yes, 3 users and 20 AI credits a month $12 per user a month Test management with 100 AI credits a month; 250 more credits cost $30 kualitee.com/pricing

Qase and Kualitee are test management tools. Their AI credits generate written tests, which a person or another tool then runs. The other 13 run tests in a browser.

The six units of AI testing tool pricing#

Per step. A step is one action in a test, such as a click or a check of what the page shows. You pay for each step a test runs, so a long test costs more than a short one. Momentic charges this way.

Per test run. A test run is one test played from start to finish, once. You pay per run whatever the test's length, sometimes with a time limit. FlowQA and QA Wolf charge this way.

Credits or tokens. You buy a monthly amount and the tool draws it down. What a credit buys differs by vendor, and several of the 15 pricing pages do not state it. Stably, TestSprite, Functionize, KaneAI, QA.tech and mabl work this way.

Per seat. You pay for each person, or each agent, with an allowance attached. TestGrid, Qase and Kualitee charge per person, and KaneAI per agent. TestSprite's billing docs say its paid plans are billed per seat and that each paid seat adds a monthly credit allowance.

Per maintained test. You pay for the size of the suite the vendor keeps working, however often it runs. Checksum prices this way.

Flat managed fee. You pay a monthly amount for people and software together. Bug0 does this.

What AI test generation costs on each tool#

AI test generation means the tool writes the test for you, from a prompt, a goal or its own exploration of your app. As of 18 November 2026, the pricing pages state the cost of that in these terms.

  • FlowQA. Starter, Team and Business include 3, 10 and 40 explorations a month, and every test you accept from an included exploration is included. Free includes 10 accepted AI-generated tests a month. Past either allowance, an accepted test costs 10 credits, or 2 with your own model key, and a rejected test costs nothing.
  • Momentic. An AI-generated step costs $0.025, twice the price of a normal step. Steps you run in the editor are free.
  • Stably. Generation draws AI tokens from the same credits as runs, at $0.30 to $15 per million tokens depending on the model.
  • KaneAI. Each plan's credits pay for the agent's work. The page estimates Starter's 2,000 credits at about 25 test creations of about 40 steps each.
  • TestGrid. Creating a typical test uses 10 to 20 tokens, and Starter includes 5,000. The page says healing and bug summaries use no tokens.
  • QA.tech. Test creation is unlimited on every plan, and only test runs use credits.
  • Functionize Studio. Every agent task draws on one credit balance and "the cost changes with the task". The page gives no rate per test.
  • TestSprite. The page lists monthly credits for each plan and does not say what generating a test uses.
  • Checksum. The page lists no charge for generation, runs or seats. You pay for the number of workflows kept working.
  • Qase. AI credits cover generating written tests and converting manual tests to automated ones. Qase's billing docs say the cost of an action varies, and give no rate per test.
  • Kualitee. A bundle of 250 AI credits for $30 is listed as up to 250 generated tests.

How FlowQA charges#

FlowQA charges per workspace, per month, and counts cloud test runs. Each browser and device a test runs on counts as one cloud test run. Runs on your own computer are free and unlimited on every plan.

Plan Price a month Cloud test runs a month Explorations a month Members
Free $0 1,000 10 AI-generated tests a month instead 1
Starter $49 3,000 3 5
Team $149 15,000 10 15
Business $399 60,000 40 unlimited

Free covers 1 project and up to 100 saved tests. A yearly plan costs the price of ten months. Every workspace can try Team for 14 days with no card.

Credits cover what goes past a plan. One credit is 1 cent: 1,000 credits cost $10. A cloud test run past the allowance is 1 credit, and a test repair is 5 credits. With your own model key, Free's allowance rises to 50 AI-generated tests a month. Billing questions has the rest.

One small suite priced under each unit#

An illustration with round numbers, priced from the linked pricing pages on 18 November 2026; it is not a customer's bill.

Take 30 tests of about 10 steps each, run once a day for 30 days, in one browser, by one person. That is 30 × 30 = 900 test runs and 900 × 10 = 9,000 steps a month. Assume each run takes two minutes.

Tool The arithmetic A month
FlowQA 900 cloud test runs are inside Free's 1,000 $0 on Free, the answer the pricing page's estimator gives for these numbers
Stably Its page puts a typical run at about $0.05 to $0.15 of credits: 900 × $0.05 = $45 and 900 × $0.15 = $135 $45 to $135 of credits; Team includes $60 and Growth $250
Momentic 9,000 steps × $0.0125 = $112.50, plus 1,800 browser minutes × $0.005 = $9 $121.50 of usage, which is past Free's $50; Free cannot add funds, so Scale at $500
TestGrid One seat; 30 tests at 10 to 20 tokens each use 300 to 600 of the 5,000 tokens $249
QA Wolf 900 web runs × $1 $900; on weekdays only, counted the way its own estimator does, 30 × 260 ÷ 12 = 650 runs, $650
Bug0 A flat fee whatever the suite runs $2,500

On FlowQA, having the AI agent write all 30 tests in the first month goes past Free's 10 accepted tests. The other 20 cost 10 credits each, 200 credits in all, at 1 cent a credit. A team of five needs Starter at $49 for the seats, and the same 900 runs fit inside it.

This suite cannot be priced on the other nine tools from what they publish. QA.tech, TestSprite and Functionize Studio do not say what one credit buys. KaneAI's page estimates credits for writing tests and does not say what a run uses. mabl, Checksum and testRigor publish no dollar figure for a paid plan. Qase and Kualitee do not run tests.

Halve the steps in each test and Momentic's usage halves, while a per-run price stays the same. Run the suite on three screen sizes and a per-run bill triples.

Where each of the other tools is the better buy#

  • Momentic costs less than a per-run tool for a few short tests run rarely, because it charges per step, and its Free plan's $50 covers 4,000 steps a month before browser minutes. It also tests native mobile apps, which FlowQA does not.
  • Stably is the better buy if you want the tests as Playwright code, which its homepage says it writes, and its pricing page lists unlimited team members on every plan.
  • TestSprite starts at $19 a month with the first month at $0, and its Free plan includes pull-request testing on one GitHub repository.
  • Functionize Studio has a Free plan with no time limit and five parallel runs, and a paid plan at $20 a month.
  • KaneAI suits a team already on TestMu AI's cloud (formerly LambdaTest), since its Pro and Max plans include execution minutes there.
  • QA.tech suits a team of up to five that wants pull requests tested, with unlimited test creation on every plan.
  • QA Wolf is the better buy for a large team that runs few tests, because it charges per run and nothing per seat, and it sells a managed service by quote.
  • TestGrid is the better buy if you need a device lab, on your premises or hosted, alongside AI test creation.
  • Bug0 is the better buy if nobody on your team will own the tests: its page says the fee covers the engineer's hours, the runs and the infrastructure.
  • Checksum suits a large suite that runs many times a day, because it does not charge per run or per seat.
  • mabl suits a company that wants browser, mobile, API, performance and accessibility testing on one credit balance and expects a quote.
  • testRigor is free if your tests and results can be public.
  • Qase and Kualitee are the better buy if what you need is written tests in a test management tool, with nothing run in a browser.

What AI test automation costs a small team covers the two costs a pricing page does not show: usage as a suite grows, and the hours spent reading failures.

FlowQA explores your site, writes the tests, keeps each one only once it fails on a page broken the way its checks describe, and runs them after every change.

See FlowQA