Mistral Large 2Top-tier reasoning for high-complexity tasks

Mistral Large 2: Advanced AI Language Model

Introduction

Mistral Large 2 is the latest generation of Mistral AI's flagship language model, offering significant improvements in code generation, mathematics, reasoning, and multilingual support. With 123 billion parameters and a 128k context window, it's designed for high-performance, single-node inference.

Key Features

• Enhanced code generation and reasoning capabilities
• Improved instruction-following and conversational abilities
• Strong multilingual support for dozens of languages
• Advanced function calling and tool use
• 128k context window
• 123 billion parameters

Use Cases

• Complex code generation tasks
• Mathematical problem-solving
• Multilingual document processing
• Long-context applications
• Business intelligence and analysis
• Research and academic applications

Pricing

Available on la Plateforme as "mistral-large-2407". Commercial licensing required for self-deployment. Contact Mistral AI for pricing details.

Teams

Accessible through major cloud service providers:

• Google Cloud Platform (Vertex AI)
• Azure AI Studio
• Amazon Bedrock
• IBM watsonx.ai

Performance Highlights

• 84.0% accuracy on MMLU (pretrained version)
• Competitive performance on code generation benchmarks
• Improved scores on mathematical reasoning tasks
• Enhanced performance on alignment and instruction-following benchmarks