Welcome Le Chonk: Mistral Large 4 on Hugging Face
Hugging Face presents Mistral Large 4 ("Le Chonk"), a sparse expert model accepting text and images, with about 1 trillion total parameters, 49 billion active and a 1-million-token context window. It ships in FP8 and NVFP4 with same-day support in transformers, vLLM and llama.cpp.
What the article says
Hugging Face published a post on October 29, 2026 presenting Mistral Large 4, nicknamed Le Chonk: a multimodal, multilingual Mixture-of-Experts model that takes text and images as input and produces text. It has about 1 trillion total parameters, 49 billion of them active, and a 1-million-token context window.
Weights come in FP8 and NVFP4 variants, with GGUFs made by Unsloth and the llama.cpp team, under the Mistral Ecosystem Open-Weight License, an MIT derivative with limits on model-as-a-service offerings. Day-0 support covers Transformers, vLLM and llama.cpp, plus Inference Endpoints and Dell Enterprise Hub. The post details the architecture, cybersecurity and coding benchmark scores, and examples for inference and fine-tuning with TRL.
Read the official article AnnuIA profile
Summary written by AnnuIA in its own words from the official page of Hugging Face. The content and rights belong to the publisher: please read the original article.