A seat costs 14 euro a month and carries a 16 euro monthly usage limit, which is 20 euro of work at the rate somebody with no plan pays. Past the limit, work carries on and is charged per unit from a balance you top up, and every unit stays 20% cheaper on Pro and 30% cheaper on Max for as long as the plan lasts. Pay as you go is permanent: nobody is ever moved off it. Every rate is visible in the console before anything runs, and every charge points back to the call that caused it.
A plan does two things: it gives every person a monthly usage limit larger than the fee, and it makes every metered unit cheaper for as long as the plan lasts. Pay as you go stays on both lists for ever, and nobody is ever moved off it. Every one of them, that one included, carries an allowance on the free model: calls that are counted rather than charged, so there is real work to do before any money is spent.
Pay as you go
-
-
Usage only, no seat fee, for as long as you like.
Monthly minimum
€5
Usage counts towards the minimum, and you pay only the shortfall.
Per person, per day: 20 free calls on the free chat model and 200 on the free embedding one, on top of everything above.
Usage only, no seat fee, for as long as you like.
Pro
€14
€11
per seat a month, net of VAT
Monthly usage limit
€16
worth €20 without a plan
About 1,600 grounded answers, or 600 automation runs, a month.
€16
worth €20 without a plan
About 1,600 grounded answers, or 600 automation runs, a month.
every unit 20% cheaper, for as long as the plan lasts
Per person, per day: 200 free calls on the free chat model and 2,000 on the free embedding one, on top of everything above.
The standard team plan.
€11 a seat a month when a year is paid up front.€134 a seat, paid once for twelve months.
Max
€24
€19
per seat a month, net of VAT
Monthly usage limit
€25
worth €36 without a plan
About 2,500 grounded answers, or 900 automation runs, a month.
€25
worth €36 without a plan
About 2,500 grounded answers, or 900 automation runs, a month.
every unit 30% cheaper, for as long as the plan lasts
Per person, per day: 500 free calls on the free chat model and 5,000 on the free embedding one, on top of everything above.
The plan for security and compliance requirements.
€19 a seat a month when a year is paid up front.€230 a seat, paid once for twelve months.
Enterprise
-
-
Negotiated, annual.
Priced per agreement
Volume rates, dedicated models and terms we both sign.
No ceiling on the free model.
Negotiated, annual.
Pay as you go
-
-
Usage only, no fixed fee, for as long as you like.
No seat fee
You pay for what you use, at the list rate, and nobody is ever moved off it.
Per person, per day: 20 free calls on the free chat model and 200 on the free embedding one, on top of everything above.
Usage only, no fixed fee, for as long as you like.
Pro
€14
€11
a month, net of VAT
Monthly usage limit
€16
worth €20 without a plan
About 1,600 grounded answers, or 600 automation runs, a month.
€16
worth €20 without a plan
About 1,600 grounded answers, or 600 automation runs, a month.
every unit 20% cheaper, for as long as the plan lasts
Per person, per day: 200 free calls on the free chat model and 2,000 on the free embedding one, on top of everything above.
For somebody who works with it daily.
€11 a seat a month when a year is paid up front.€134 a seat, paid once for twelve months.
Max
€24
€19
a month, net of VAT
Monthly usage limit
€25
worth €36 without a plan
About 2,500 grounded answers, or 900 automation runs, a month.
€25
worth €36 without a plan
About 2,500 grounded answers, or 900 automation runs, a month.
every unit 30% cheaper, for as long as the plan lasts
Per person, per day: 500 free calls on the free chat model and 5,000 on the free embedding one, on top of everything above.
Every model including the strongest.
€19 a seat a month when a year is paid up front.€230 a seat, paid once for twelve months.
A plan buys two things: a monthly limit larger than the fee, and a cheaper rate on every unit for as long as the plan lasts. Below the seat fee in usage, pay as you go costs less, and it is on this page rather than hidden.
Every figure on this page is net. Value added tax is added at your own country's rate at the checkout, and a business with a valid identifier in the Union is invoiced with the tax reverse charged.
Each unit is counted where it happens and priced from the catalogue your organization sees.
Input and output tokens, per model, at the rate published for that model.
Per query, when a conversation or an agent searches the web.
Per second of sandbox time, counted while your code actually runs.
Per execution of a published function, whether it was called by a URL, an agent or a schedule.
Stored data per day, bytes sent out, and requests served by object storage.
Per document operation, and per second of processor and memory a site uses to answer.
Set the sliders to your own shape of work. The estimate uses the launch rates from the catalogue, states every assumption it makes, and is worth exactly as much as those assumptions.
Estimated monthly spend
per person a month
The same month, priced three ways Monthly
Pay as you go
On Pro
On Max
These are the platform defaults published in the catalogue. Your organization sees its own rates in the console, and nothing is charged until a call actually runs.
The figures are an estimate for guidance only. They are not a quote and not a commercial offer, and the real bill is decided by real usage.
Add credit when you want to. The balance is the organization's, and every service spends from the same one.
A request that would exceed the balance or a limit is refused before it runs, with the reason in the error, rather than appearing on an invoice later.
The worst case cost is held while the request is upstream and the real amount is settled when it returns, so concurrent traffic cannot spend the same balance twice.
Everything in euro, with usage broken down per day, per model and per key.
A daily or monthly ceiling on everything the organization spends.
An integration gets its own ceiling and is refused at it, so one script cannot spend the department's month.
The platform tells somebody when a balance is running low rather than only when it stops working.
Create an account, look at the rates for the models you would use, and decide afterwards.