Mistral API

The Mistral API in Europe, and its alternatives

Mistral AI is a French company, so the question is not whether Mistral is European. It is where a given request runs, what is kept, and what else sits next to Mistral in your stack. Mistral's default endpoint does not commit to an inference location; its EU region does, at 1.1 times the list price. On Frontière AI, Mistral Small 3.2 24B, Mistral Large 4 (preview, 512K context) and Mistral Small 4 119B run in Europe and are labeled EU sovereign, behind the same OpenAI-compatible API as the DeepSeek, Qwen and GLM models hosted in Europe, prepaid, prices VAT included. When Mistral serves the model, Mistral's own data rules still apply; this page says which models that concerns, and lists EU-hosted alternatives to Mistral.

Where are Mistral models hosted on Frontière AI?

In Europe, by Scaleway (France) for Mistral Small 3.2 24B and Mistral AI (France) for Mistral Large 4 and Mistral Small 4 119B. The provider is named on each model before you call it, and each response names it again in the x-routed-provider header.

Mistral models on Frontière AI

Every Mistral model we serve, with the provider that serves it and its jurisdiction label — taken from our live catalog when this page is built.

Prices margin and VAT included — the amount debited from your prepaid balance. Tool calling measured with one real call per model.

Mistral's own API, side by side

What Mistral AI's own documents say, checked on 2026-10-13, next to the same questions on Frontière AI. We quote; we do not interpret on their behalf.

Mistral AI sources, checked on 2026-10-13: Mistral AI — commercial terms of service · Mistral AI — privacy policy · Mistral AI — regional inference · Mistral AI — training on your data (help centre) · Mistral AI — pricing · Mistral AI — Mistral Large 4 (Public Preview)

EU-hosted alternatives to Mistral

Looking for a Mistral alternative without leaving Europe? These models from other makers are served by European providers on Frontière AI and labeled EU sovereign, cheapest on input first. They sit behind the same API as the Mistral models, so switching is a change of model ID.

Prices per 100k tokens (in / out), margin and VAT included — taken from the live catalog when this page is built.

Open weights, closed weights

Mistral publishes the weights of some models under Apache 2.0 and keeps others closed or in preview. That decides who can serve them: an open-weight Mistral model can run at another European operator, while a closed or preview model is served by Mistral itself. Each model page on Frontière AI states which applies, with its licence and the official model card.

Switching takes one line

Frontière AI speaks the OpenAI format. With the OpenAI SDK, only the base URL, the key and the model ID change:

from openai import OpenAI

client = OpenAI(
    base_url="https://getfrontiereai.eu/api/v1",
    api_key=os.environ["FRONTIERE_KEY"],
)
reply = client.chat.completions.create(
    model="mistral-small-3.2-24b",  # EU sovereign: served in Europe, labeled before you call it
    messages=[{"role": "user", "content": "Bonjour"}],
)

FAQ