🌬️

Mistral

Mistral is the model family from Mistral AI, a French AI lab, spanning the flagship Mistral Large through the open-weight Mixtral mixture-of-experts models. It ranks #11 in usage share at roughly 10.4%. Mistral is known for efficient, strong-performing models available both as open weights and as an API, and is especially popular in Europe for data-sovereignty-conscious deployments where data needs to legally stay within the EU.

Quick facts
Made by: Mistral AI (France)
Access: Paid API (free trial credits) + open-weight self-hostable models (Mixtral, Mistral Small)
Context window: ~32K–128K tokens depending on model
Strengths: Efficient inference, strong multilingual (especially European languages) performance, EU data hosting
Primary use case: EU-based enterprise deployments, multilingual apps, cost-efficient self-hosted inference
Jump to: ExampleGetting startedBest for

Example

Call the Mistral API directly with the official mistralai Python package.

from mistralai import Mistral

client = Mistral(api_key="YOUR_API_KEY")

response = client.chat.complete(
    model="mistral-large-latest",
    messages=[{"role": "user", "content": "Translate 'good morning' into French, German, and Spanish."}]
)

print(response.choices[0].message.content)

Getting started

Sign up on La Plateforme (Mistral's developer console) for an API key, or skip the API entirely and self-host an open Mixtral model with Ollama.

# 1. sign up at console.mistral.ai for an API key

# 2. install the SDK
pip install mistralai

# 3. or self-host the open-weight model instead
ollama run mixtral
Best for: European companies and regulated industries that need AI processing to stay within EU jurisdiction, plus teams that want an efficient open-weight model (Mixtral) they can self-host at lower cost than proprietary alternatives.