mistralai/Mistral-Large-Instruct-2411

Hugging Face
TEXT GENERATIONConcurrency Cost:4Model Size:123BQuant:FP8Ctx Length:32kPublished:Nov 14, 2024License:mrlArchitecture:Transformer0.3K Warm

Mistral-Large-Instruct-2411 is an advanced 123 billion parameter dense Large Language Model developed by Mistral AI, featuring a 128k token context window. It excels in reasoning, knowledge, and coding across dozens of languages, including Python, Java, and C++. This model is particularly optimized for agentic capabilities with native function calling, robust context adherence for RAG, and improved system prompt handling, making it suitable for complex, multi-turn conversational AI and automated task execution.

Loading preview...

Mistral-Large-Instruct-2411 Overview

Mistral-Large-Instruct-2411 is an advanced 123 billion parameter dense Large Language Model from Mistral AI, building upon Mistral-Large-Instruct-2407. It features a substantial 128k token context window and significant enhancements in long context adherence, function calling, and system prompt support. This model is designed for sophisticated AI applications requiring high-level reasoning and precise instruction following.

Key Capabilities

  • Multi-lingual Proficiency: Supports dozens of languages, including English, French, German, Spanish, Italian, Chinese, Japanese, and Korean.
  • Advanced Coding: Trained on over 80 programming languages, such as Python, Java, C, C++, JavaScript, Bash, Swift, and Fortran.
  • Agent-Centric Design: Offers best-in-class agentic capabilities with native function calling and reliable JSON outputting.
  • Superior Reasoning: Demonstrates state-of-the-art mathematical and general reasoning abilities.
  • Robust Context Adherence: Ensures strong performance in RAG (Retrieval Augmented Generation) and other large context applications.
  • Enhanced System Prompt Support: Provides more reliable adherence to system prompts, recommending clear purpose outlines for optimal results.

Good For

  • Developing complex AI agents requiring precise function calling and structured outputs.
  • Applications demanding high-quality reasoning and mathematical problem-solving.
  • Multilingual chatbots and content generation across a wide array of languages.
  • Code generation, analysis, and translation in diverse programming environments.
  • RAG systems and applications that benefit from a very large context window and strong context adherence.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p