TestingBudget

Build

Build tokens

The AI model tokens an AI-assisted tool consumes while authoring the suite — a real cost line, separate from people.

Unit
tokens, one-time
Where you enter it
Computed from the token sub-fields on the Build tab; shown when AI tokens are toggled on.

What it means

Tools that use large language models to generate or repair test cases consume tokens as they do so, and those tokens are metered. Build tokens is the sum of the per-case token figures across the suite plus any one-time steps, priced at the cost per million tokens set in Settings.

It is shown only when the AI-tokens toggle is on, and only for automation vendors; manual approaches have no token line.

Why it matters to the cost

At around 150,000 tokens per average case and $3 per million, a 300-case suite costs roughly $135 in tokens to build — small next to labour, but it scales with the suite and with how much the tool re-generates during repair. The model reports token spend as a share of total cost as a sanity check on any AI tool's economics.

Typical values

AI-native tools: around 150,000 tokens per average case authored. Other categories: none.

Typical values describe a category of tool — AI-native, codeless, script-based, homegrown, manual — never a named product. Why we do not price named products.

Where it appears in the model

Every name is a link to its definition. See the whole model.

Try it with your numbers

Build tokens is a row in the calculator. Change it and every total on the page moves with it — free, in your browser, nothing to install.

See how build tokens impacts the calculation of test automation costs

Last reviewed 2026-09-15. All terms.