Hugging FaceFrom the builder
Welcome Le Chonk: Mistral Large 4 on Hugging Face
In a 29 October 2026 post, Hugging Face covers Mistral Large 4 (aka Le Chonk): a multimodal, multilingual MoE with about 1T total parameters, about 49B active, and a 1M-token context, shipping official FP8 and NVFP4 weights plus GGUFs from Unsloth and the llama.cpp team under the Mistral Ecosystem Open-Weight License.
The post highlights cybersecurity, coding agents and multimodality, describes sparse attention with a Pixtral vision encoder, and walks through Transformers, vLLM, llama.cpp, TRL fine-tuning, Inference Endpoints and Dell Enterprise Hub deploy paths.
Mistral Large 4 (aka Le Chonk) is a multimodal, multilingual, Mixture-of-Experts (MoE) model from Mistral. It has about 1T total, 49B active parameter, and a context window of 1M tokens.
Saved . A shortened excerpt from the official blog; the original publisher retains authorship.