The model
Token cost
AI model tokens consumed, divided by a million, multiplied by the price per million tokens set in Settings.
- Unit
- money
- Where you enter it
- Token sub-fields in the tables, shown when AI tokens are toggled on; the price is in Settings.
What it means
AI-assisted tools consume metered model tokens while authoring, triaging, healing and sometimes executing tests. Each token figure in the model is priced at the cost per million tokens in Settings and added to the labour cost of its phase.
Tokens are consumed by what the machine does regardless of who is watching, so they are never scaled by human effort. Manual approaches have no tokens.
Why it matters to the cost
For most suites tokens are a small share of cost, but they scale with suite size and with how much a tool re-generates during repair. The model reports token spend as a percentage of total cost as a sanity check: if it looks implausible, the authoring assumptions usually need revisiting.
Typical values
Around $3 per million tokens at the default; a few hundred dollars a year for a 300-case suite on an AI-native tool.
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
Build — one-off
- Total test cases
- basic + average + complex Test case complexity tiers
- Build effort
- Set up, install and configure + Scoping the suite + Training time per person × People trained
- + Σ Test case complexity tiers × Build time per test case for each tier, as Scratch vs migration
- Build cost
- Build effort × Effective hourly rate + Build tokens priced as Token cost
Run — every release
- tests broken per run
- Total test cases × Test breakage rate ÷ Runs per year — found here, repaired under Maintain
- flaky per run
- Total test cases × Flaky failures per run
- Failures per run
- Defects found per run + tests broken per run + flaky per run — derived, never asked
- Run effort
- Prepare and start a run + Failures per run × Triage one failure + Defects found per run × Write up one defect + Report and sign off — an automation tool
- Run effort
- Prepare and start a run + Total test cases × Execution time per case + Defects found per run × Write up one defect + Report and sign off — a manual approach — the tester is the run
- Machine time per run
- Total test cases × Execution time per case — automation only; duration, not labour
- Run cost per year
- Run effort per run × Runs per year × Effective hourly rate + Token cost
Maintain — continuously
- cases to repair
- Total test cases × Test breakage rate
- cases healed
- Total test cases × Self-healed cases per year
- Maintenance effort per year
- cases to repair × Repair time per case + cases healed × Review one heal + Platform upkeep per year + New cases added per year × Build a new case
Turning effort into cost
- Effective hourly rate
- loaded cost per FTE per year ÷ productive hours per FTE per year
- Labour cost
- hours × Effective hourly rate
- Token cost
- tokens ÷ 1,000,000 × cost per million tokens
Every name is a link to its definition. See the whole model.
Try it with your numbers
The calculator works token cost out from the rows it is built on. Change those and every total on the page moves with them — free, in your browser, nothing to install.
See how token cost impacts the calculation of test automation costsLast reviewed 2026-09-15. All terms.