M

Mistral

⭐⭐⭐⭐⭐ 4.6/5

European AI company with powerful open source models and innovative MoE architecture

open-sourcellmeuropeancoding
Visit Website

Developer

Mistral AI

Released

Sep 2023

Users

30M+

API

Available

📊 Rating Breakdown

Functionality 4.7/5

Mistral 7B, Mixtral 8x7B, Codestral, reasoning, coding

Ease of Use 4.5/5

Good API, Le Chat interface, open source options

Performance 4.6/5

Excellent for the size, very efficient

Value for Money 4.8/5

Open source, API pricing is very competitive

Ecosystem 4.6/5

Growing, strong European presence

Privacy & Security 4.7/5

EU-based, GDPR compliant, self-hostable

Support 4.4/5

Good docs, community

Innovation 4.7/5

Mixture of Experts innovation, efficiency leadership

✅ Pros

  • European AI leader
  • Excellent efficiency
  • Innovative MoE architecture
  • Good open source options
  • GDPR compliant

❌ Cons

  • Less capable than top closed models
  • Smaller ecosystem
  • Technical for new users
  • Less multimodal
  • Growing but still smaller community

💰 Pricing

Very competitive pricing

Visit Website

🧠 Available Models

Mistral Large 3

Proprietary

Mistral's flagship multimodal model with 675B total parameters (MoE). Agentic capabilities, 256K context, function calling, JSON mode. Open-weight.

Parameters: 675B total (MoE)
Context: 256K tokens
📝 text 🖼️ image
✅ Multilingual ✅ Agentic tasks ✅ Function calling ✅ Open-weight
Input: $2/1M tokens
Output: $6/1M tokens
🔌 API Availablevia Mistral AI API, Azure AI, Vertex AI, Bedrock

Pixtral Large

Proprietary

Mistral's vision-language model built on Mistral Large. Native multimodal understanding with strong document analysis.

Parameters: 124B (MoE)
Context: 128K tokens
📝 text 🖼️ image 📄 document
✅ Document understanding ✅ OCR ✅ Chart analysis ✅ Multimodal
Input: $2/1M tokens
Output: $6/1M tokens
🔌 API Availablevia Mistral AI API

Mistral Small

Open Source

Efficient low-latency model optimized for simple tasks, classification, and customer support. Best cost-performance ratio.

Parameters: 22B (dense)
Context: 128K tokens
📝 text
✅ Fast ✅ Low cost ✅ Efficient ✅ Ideal for simple tasks
Input: $0.1/1M tokens
Output: $0.3/1M tokens
🔌 API Availablevia Mistral AI API

Codestral

Open Source

Mistral's specialized code generation model. Fill-in-the-middle, advanced code completion, and code reasoning.

Parameters: 22B
Context: 256K tokens
📝 text 💻 code
✅ Code generation ✅ Fill-in-middle ✅ Multi-language coding
Input: $0.2/1M tokens
Output: $0.6/1M tokens
🔌 API Availablevia Mistral AI API

Mistral Nemo

Open Source

12B model developed with NVIDIA, optimized for efficiency and multilingual tasks.

Parameters: 12B
Context: 128K tokens
📝 text
✅ Lightweight ✅ Multilingual ✅ Efficient
Input: $0.15/1M tokens
Output: $0.15/1M tokens
🔌 API Availablevia Mistral AI API, NVIDIA NIM

API Quickstart Guide

Visit the official API documentation to get started with this tool.

🧠 Available Models

Mistral Large 3

Proprietary

Mistral's flagship multimodal model with 675B total parameters (MoE). Agentic capabilities, 256K context, function calling, JSON mode. Open-weight.

Parameters: 675B total (MoE)
Context: 256K tokens
📝 text 🖼️ image
✅ Multilingual ✅ Agentic tasks ✅ Function calling ✅ Open-weight
Input: $2/1M tokens
Output: $6/1M tokens
🔌 API Availablevia Mistral AI API, Azure AI, Vertex AI, Bedrock

Pixtral Large

Proprietary

Mistral's vision-language model built on Mistral Large. Native multimodal understanding with strong document analysis.

Parameters: 124B (MoE)
Context: 128K tokens
📝 text 🖼️ image 📄 document
✅ Document understanding ✅ OCR ✅ Chart analysis ✅ Multimodal
Input: $2/1M tokens
Output: $6/1M tokens
🔌 API Availablevia Mistral AI API

Mistral Small

Open Source

Efficient low-latency model optimized for simple tasks, classification, and customer support. Best cost-performance ratio.

Parameters: 22B (dense)
Context: 128K tokens
📝 text
✅ Fast ✅ Low cost ✅ Efficient ✅ Ideal for simple tasks
Input: $0.1/1M tokens
Output: $0.3/1M tokens
🔌 API Availablevia Mistral AI API

Codestral

Open Source

Mistral's specialized code generation model. Fill-in-the-middle, advanced code completion, and code reasoning.

Parameters: 22B
Context: 256K tokens
📝 text 💻 code
✅ Code generation ✅ Fill-in-middle ✅ Multi-language coding
Input: $0.2/1M tokens
Output: $0.6/1M tokens
🔌 API Availablevia Mistral AI API

Mistral Nemo

Open Source

12B model developed with NVIDIA, optimized for efficiency and multilingual tasks.

Parameters: 12B
Context: 128K tokens
📝 text
✅ Lightweight ✅ Multilingual ✅ Efficient
Input: $0.15/1M tokens
Output: $0.15/1M tokens
🔌 API Availablevia Mistral AI API, NVIDIA NIM

API Quickstart Guide

Visit the official API documentation to get started with this tool.

📜 Version History

Sep 2025 Mistral 3.5

Mistral 3.5 Suite

Enhanced Mistral 3.5 suite with agentic capabilities, enterprise deployment, and custom fine-tuning platform.

Sep 2023 Mistral 7B

Mistral 7B Launch

Mistral AI released Mistral 7B, an open-source model that outperformed larger models on many benchmarks.

Nov 2024 Mistral Small

Mistral Small & Pixtral

Released optimized Mistral Small for efficient deployment and Pixtral 12B for multimodal tasks.

Mar 2025 Mistral 3.0

Mistral 3.0 Launch

Next-gen Mistral 3.0 with native multimodality, 256K context, and industry-leading efficiency.

Jul 2024 Mistral Large 2

Mistral Large 2

Released Mistral Large 2 with 123B parameters and significantly improved multilingual capabilities.

Feb 2026 Mistral 4.0

Mistral 4.0 Release

Released Mistral 4.0 with breakthrough reasoning, massive context, and autonomous agent orchestration.

Feb 2024 Mistral Large

Mistral Large & le Chat

Released Mistral Large (top-tier model) and launched le Chat, Mistral's conversational AI assistant.

Dec 2023 Mistral Medium

Mistral Medium & API

Released Mistral Medium via API and launched Mistral's cloud platform for developers.

Apr 2024 Mixtral 8x22B

Mixtral 8x22B MoE

Released Mixtral 8x22B, a massive Mixture-of-Experts model with exceptional performance and efficiency.

📄 Subpages

📖

How to Use
💰

Pricing
🔄

Alternatives
🎁

Free
📋

Changelog