Mistral AI Team Base Fee, API Rates & the Real Bill 2026 Guide

Le Chat seats read simply. Free is $0, Pro is $14.99 a user, Team is $24.99 a user. The surprise sits on the Team plan, which adds a flat $50 a month base fee on top of every seat. So a five-person Team pays $124.95 in seats and another $50 in base fee, a line the seat card never leads with.

The API is a separate meter entirely. Mistral Large runs about $2 per million input tokens and $6 per million output, while the open Mistral Small 4 is far cheaper at $0.15 in and $0.60 out. Developers scaling the hosted flagship report the per-token cost adding up, so which model you route to sets the API bill more than volume does.

Document work has its own line. OCR processing costs $4 per 1,000 pages on demand, or $2 per 1,000 through the Batch API. None of that is on the Le Chat plan card. The full seat, API and OCR rates sit on the Mistral AI pricing page, so a real budget is three separate meters added together, not one seat number.

The Free tier is unusually strong. It gives web and mobile access to Le Chat with the current SOTA models, image generation, search, and over 40 enterprise connectors, all at $0. Because the models are the same ones the paid tiers run, many people can genuinely live on it for a while before caps bite.

The wall is limits, not capability. Free caps messages, searches, image generations and think-mode uses, so heavy daily work hits the ceiling. That is where Pro at $14.99 lifts the caps and opens all-day coding across the CLI, IDE and web. Before you pay, weigh Mistral against the cheaper US assistants, since the Mistral AI alternatives page shows a Gemini seat starts lower while a ChatGPT one sits nearby.

Mistral runs one genuine discount, which is rare in this category. Verified students pay $5.99 a month for Pro instead of $14.99, for up to twelve months. That is the full Pro feature set at less than half price, so if you can verify as a student there is no reason to pay the standard rate.

Beyond that, there is no published nonprofit or startup program as of July 2026. The other savings are structural. Route API work to the cheaper open models, batch OCR to halve the per-page rate, and negotiate the Team base fee and API into an Enterprise contract. That negotiated lane is what the tactics below cover.

The consumer seats are fixed. Nobody discounts a $14.99 Pro, and the student rate is the only self-serve break. The conversation opens at Team and Enterprise, where the $50 base fee, seat count and hosted-model rates all become negotiable at volume.

Two things carry the weight here, and both work because Mistral is a challenger fighting for enterprise footholds against pricier US labs.

The consumer seats never change with the calendar, so timing is only a factor for Team and Enterprise deals. Mistral is a challenger chasing enterprise wins, which works in a buyer's favor. A competitive deal it can book near a quarter close often prices better. A hungry sales team is also more willing to waive the base fee to land a logo.

Aim requests where Mistral can move. Consumer seats and the student rate are fixed; the room is in the Team base fee, Enterprise seats and hosted-model rates.

Fill the fields and the draft below writes itself, folding in live rival prices from our verified catalog. Direct it to your Mistral account manager or the enterprise sales form. Lead with the numbers. Give your seat count and monthly token volume. Name a cheaper assistant beside it, ask to waive the Team base fee and lower the seat rate, and set a date to decide by.

Each of these follows from the way Mistral stacks seats, a base fee and metered usage, and all are avoidable up front.

Mistral's paid pitch is EU hosting and data residency, so leverage means naming the assistants that do similar work for less and making it defend the premium. These three are priced from our verified catalog. You need not switch. You need a rival seat with a number attached, run through a real task, so the compliance premium is argued rather than assumed.