Mistral

Mistral provides efficient, open-source models optimized for production — specialized for code generation with the Codestral model and strong general-purpose reasoning.

Why Vibe Coders Use It

• Codestral — dedicated code generation model, highly specialized for programming
• Efficient architecture — smaller models deliver impressive performance
• Open-source foundation — full control with Apache 2.0 licensed models
• Available everywhere — cloud API (La Plateforme), Azure, AWS Bedrock, self-hosted
• Cost-effective — competitive pricing with strong quality

Key Specs

Getting Started

Option 1: Via Mistral Cloud (La Plateforme)

npm install @ai-sdk/mistral

import { mistral } from '@ai-sdk/mistral';
import { generateText } from 'ai';

const { text } = await generateText({
  model: mistral('mistral-large-latest'),
  prompt: 'Write a React component that manages form state with validation',
});

console.log(text);

Option 2: Using Codestral (Code Specialist)

import { mistral } from '@ai-sdk/mistral';
import { generateText } from 'ai';

// Use Codestral for specialized code generation
const { text } = await generateText({
  model: mistral('codestral-latest'),
  prompt: 'Generate SQL to find the top 10 users by activity',
});

console.log(text);

Option 3: Via Azure

npm install @ai-sdk/mistral

import { mistral } from '@ai-sdk/mistral';
import { generateText } from 'ai';

const { text } = await generateText({
  model: mistral('mistral-7b-instruct'),
  prompt: 'Explain how JavaScript closures work',
});

Option 4: Self-Hosted with Ollama

ollama run mistral

# Or for Codestral
ollama run codestral

// Query your local server
const response = await fetch('http://localhost:11434/api/generate', {
  method: 'POST',
  body: JSON.stringify({
    model: 'codestral',
    prompt: 'Write a function to calculate Fibonacci numbers',
  }),
});

Streaming Chat Example

import { mistral } from '@ai-sdk/mistral';
import { streamText } from 'ai';

export async function POST(req: Request) {
  const { messages } = await req.json();

  const result = streamText({
    model: mistral('mistral-large-latest'),
    messages,
  });

  return result.toDataStreamResponse();
}

When to Use Mistral vs. Alternatives

Use Mistral (especially Codestral) when you want open-source code generation. Use Claude for the strongest reasoning. Use GPT for general-purpose reliability and speed.

Available Models

• Codestral — Specialized for code generation, highly optimized
• Mistral Large — General-purpose, powerful reasoning
• Mistral 7B — Lightweight, efficient
• Mixtral 8x7B — Mixture of Experts architecture

Resources

• Mistral AI Official
• La Plateforme Console
• Mistral Documentation
• Ollama Mistral Models
• AI SDK Mistral Provider
• See the full Mistral profile on LLMReference →