Find out what your team should be paying for AI.
Tell us which AI tools and models your team pays for. You get a written report: what you pay now, what the same work costs on the alternatives, and what to change.
Two sizes
- Up to 5 tools or models
- Your current monthly bill, line by line
- The same usage priced on up to 3 alternatives each
- A one-page summary with the changes worth making
- Up to 15 tools or models, per team or project
- Seat plans, usage allowances and overage checked against how you use them
- API workloads priced on current vendor rates, with cache and batch options
- Retirements and announced price changes that affect you in the next 6 months
- A 45-minute call to walk through it
An example, from today’s data
A five-person engineering team choosing a coding assistant. These are the published plan prices for five seats, before any extra usage, read on each vendor’s page.
| Tool and plan | 5 seats / month | What each seat includes |
|---|---|---|
| CursorTeams (Premium seat) | $150.00 | Each seat matches the individual plan of the same tier; seat tiers can be mixed. |
| GitHub CopilotBusiness | $95.00 | 1,900 AI credits per user a month, pooled across the organization. |
| Claude CodeTeam (Standard seat) | $125.00 | More usage than Pro per seat (no number); teams of 2 to 150. |
The gap between the cheapest and the dearest of these is $660 a year for one team, before usage. The audit does this for your real tools, your real usage and your contract terms, and for API work: today the cheapest verified model for a typical coding job is Ling 3.0 Tiny at $15.60 per 1,000 tasks (how that is worked out).
Keep it current: AI spend watch
The audit, kept current. Each month we re-price your stack on the latest vendor prices and tell you what changed, what is about to change, and what it does to your bill.
- Up to 10 tools or models
- A monthly report: your bill re-priced, with every change since last month
- An email within one working day when a price, plan or retirement that affects you is recorded
- Cancel any month
- Up to 40 tools or models, across teams
- Everything in Team, split per team or project
- API workloads re-priced on current rates each month
- A call each quarter to walk through it
How it works
- 01
You send the list
The tools, plans and models you pay for, roughly how much you use them, and what the team does with them. An email is enough; no access to your systems is needed.
- 02
We price it
Every line is priced from the vendor’s own current pages, the same sourced and dated data as this site, including limits the vendor does not publish, stated as unknown.
- 03
You get the report
What you pay, what the same work costs elsewhere, what is about to change (retirements, announced price moves), and the changes worth making, with the saving for each.
What it is not
- It is not a recommendation paid for by a vendor. No vendor pays us, and we take no commission on anything the report suggests.
- It does not judge output quality for your work: it compares cost, limits and published facts. Where quality evidence exists we cite it, with its source.
- It is not financial or legal advice.
Prices are in euros and exclude VAT where it applies. You pay by invoice after accepting a written quote. If the report is not delivered in the stated time, you get your money back. Terms of sale and refunds
Never for sale
- Rank and peer-board membershipdecided by the published method and its evidence only.
- Eligibility and confidenceevidence coverage is reported as found; missing data stays unknown.
- Publish timing and review speedthe catalogue is published when its checks pass; no one can pay to be reviewed sooner.
- Placementno sponsored rows, no promoted cards, no paid badges anywhere in the catalogue.
Enforced by build tests and stated in full in the commercial policy. A customer of any plan is listed, checked and ranked exactly like everyone else.