Mistral Vibe Pro API credit, and the two bills behind it
By Christophe Paka · Facts checked on 22 October 2026
Mistral, a French AI company, sends two separate bills: one for its chat app, Vibe, and one for its API, the route your code uses. This page shows both, and the Mistral Vibe Pro API credit that links them.
"I am struggling to identify what API quota is included in Le Chat Pro plan," one reader wrote before paying. Le Chat is the old name of Vibe. Another, a subscriber for half a year, suddenly got mails saying he was running out of monthly quota. Both were reading one bill while two were running.
The figure below puts the two bills side by side.
How to read this figure
Start with the left meter. Studio is Mistral's API, the door your code uses. Chat models bill by the token, a small piece of text the model reads or writes. Other services use their own unit: pages for reading scans, characters or minutes for voice. If you only chat in the app, you never count tokens.
Read the four API prices. Each line gives input, then output, per million tokens. Medium 3.5 costs $1.5 and $7.5, Large 4 in preview $1.36 and $4.18, Large 3 $0.5 and $1.5, and Small 4 $0.15 and $0.6. Output costs more on every one.
Move to the right meter. Vibe is the chat app, billed per seat. Free costs $0 and lists no API credit. Pro costs $14.99 with a $15 API credit, both monthly. Team costs $24.99 per user. Beside Team the figure writes "$50 min": Mistral's card shows $50 a month next to Users and does not say what it is. It may be a minimum of two seats, but that is a guess. Enterprise is custom.
Budget Team safely. Because Mistral does not say what the $50 is, the book's advice is simple: budget $50 a month for Team, and ask Mistral sales before you buy one seat. The post on the four Vibe plans compares all four.
Find the one bridge. Pro's $15 credit is the only thing that crosses. It is spent at the same rates as any Studio bill. Past it, you pay for more use at the API rate. A seat buys no other API use, and API spend never buys a seat.
Add the tax. The listed prices exclude VAT, the sales tax added in Europe. Buyers in the EU pay more than the figure shows, so add your VAT rate before you compare plans.
Key takeaways
- Keep two budget lines, one for Vibe seats and one for the API, and never let one pay for the other.
- Treat Pro's API credit as a small allowance spent at API rates, not as unlimited use.
- Set a spending cap below your prepaid credit before your code sends real traffic.
Worked example: one small firm's month on both meters
Rivière & Co, the small firm in the book, has six Team seats. It also runs an email draft job on Small 4 through the API: 6,600 drafts in a month.
- Seats: 6 times $24.99 is $149.94 for the month, before VAT.
- API: the 6,600 drafts on Small 4 cost $3.07 for the month.
- Total: $3.07 plus $149.94 is $153.01 for the month.
Result: The seats, not the tokens, are most of the bill. Budget each line on its own meter, and watch the seat count as closely as the token count.
Questions readers ask
What API quota comes with Mistral Vibe Pro?
A $15 credit each month, spent at the same rates as any Studio bill. Past it, you pay for more use at the API rate. Pro gives no other API use, and the Free plan lists no API credit.
How far does the Pro API credit go?
It depends on the model. In the book's example, a small firm sends 6,600 email drafts a month on Small 4 for $3.07, well inside $15. The same month on Medium 3.5 costs $34.65, more than twice the credit. Past the credit, use bills at the API rate.
Is Vibe billing separate from the Mistral API?
Yes. They are two meters with two invoices. A Vibe seat is billed per user each month; the API bills your code by the token, page or minute. Only Pro's API credit crosses between them. A spending cap set for your organization covers both together.
Can I cap my Mistral bill?
Yes. Set a spending cap per organization or per workspace in the Admin Panel. When the organization reaches it, API access can be suspended until the next month begins or an admin raises the limit. Requests then fail, and nothing more is charged.
Sources: Pricing (Mistral AI); Pricing (Mistral docs); Usage and limits (Mistral docs).
This figure is from Mistral Made Crystal Clear, chapter 5.
The chapter builds the idea step by step, with recipes you can follow, a Try it to check yourself, and every source.
Get the book
