Model pricing

Compare input and output costs. Every price shows the provider rate and our service fee.

USD per million tokens, before tax.
ModelInputOutput
gpt-5.6-sol
$4.40Before tax
Provider rate
$4.00
Service fee (10%)
$0.40
$22.00Before tax
Provider rate
$20.00
Service fee (10%)
$2.00
gpt-5.6-terra
$2.20Before tax
Provider rate
$2.00
Service fee (10%)
$0.20
$13.20Before tax
Provider rate
$12.00
Service fee (10%)
$1.20
o3
$2.20Before tax
Provider rate
$2.00
Service fee (10%)
$0.20
$8.80Before tax
Provider rate
$8.00
Service fee (10%)
$0.80

Rates include the provider cost and a 10% service fee. Prices shown are before tax. After sign-in, your billing profile determines any applicable tax before the final charge. Verified 12 September 2026

How prepaid billing works

Funds are reserved before a request runs, then settled against actual usage.

1  Reserve

A request starts only when its reservation fits your available credit and spending limits.

2  Process

Input and output tokens are taken from the provider's usage report.

3  Settle

Actual usage replaces the reservation. The unused amount returns to your balance.

Pricing questions

What if a stream cuts out halfway?

Charges follow confirmed provider usage. If usage cannot be confirmed, the reservation stays pending while it is checked. A disconnected stream does not by itself release the reservation.

Can my balance go negative?

An enabled overdraft can allow a negative available balance. Your organization's credit and spending limits still apply.

Does unused balance expire?

No. Funded balance remains available until it is used or refunded under the terms.

Is there a free tier?

No. Usage is prepaid from the first request.