Mistral Large 4 (also known as Le Chonk) is a multimodal, multilingual Mixture-of-Experts (MoE) model from Mistral. It takes text and images as inputs and generates text. The model features approximately 1T total parameters with 49B active parameters, along with a 1M token context window.
The model is released in two official versions:
In addition, GGUF versions are available, built by Unsloth and the llama.cpp team. The model is released under the Mistral Ecosystem Open-Weight License, an MIT-derivative license that includes restrictions on Model-as-a-Service usage.
Mistral Large 4 demonstrates particularly strong performance in cybersecurity, coding agents, and multimodal tasks. In this post, we’ll dive into how the model works, its capabilities, architecture, and more!