# Welcome Le Chonk: Mistral Large 4 on Hugging Face

DevFeed: [Welcome Le Chonk: Mistral Large 4 on Hugging Face](<https://devfeed.tech/articles/welcome-le-chonk-mistral-large-4-on-hugging-face-107127.md>)

Original publisher: [Read original article](<https://huggingface.co/blog/mistral-large4>)

Author: merve; Alvaro Bartolome; Rémi Ouazan Reboul; Cyril Vallez; Quentin Gallouédec; Enrique Hernández Calabrés; Juan Julián; Xuan-Son Nguyen; Pedro Cuenca; ben burtenshaw

Published: 2026-10-29T00:00:00Z

Content type: article

Language: en

Sources: [Hugging Face](<https://devfeed.tech/sources/hugging-face-blog.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Generative AI](<https://devfeed.tech/topics/generative-ai.md>), [LLM security](<https://devfeed.tech/topics/llm-security.md>), [llama.cpp](<https://devfeed.tech/topics/llama-cpp.md>), [vllm](<https://devfeed.tech/topics/vllm.md>), [inference-endpoints](<https://devfeed.tech/topics/inference-endpoints.md>), [Self Hosted AI Tools](<https://devfeed.tech/topics/self-hosted-ai-tools.md>), [On-device AI](<https://devfeed.tech/topics/on-device-ai.md>), [peft](<https://devfeed.tech/topics/peft.md>), [lora](<https://devfeed.tech/topics/lora.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-workflows](<https://devfeed.tech/tags/agentic-workflows.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-servers](<https://devfeed.tech/tags/ai-servers.md>), [amd](<https://devfeed.tech/tags/amd.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [api](<https://devfeed.tech/tags/api.md>), [app](<https://devfeed.tech/tags/app.md>), [apple](<https://devfeed.tech/tags/apple.md>), [apple-silicon](<https://devfeed.tech/tags/apple-silicon.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [artificial-analysis](<https://devfeed.tech/tags/artificial-analysis.md>), [hugging-face](<https://devfeed.tech/tags/hugging-face.md>), [inference](<https://devfeed.tech/tags/inference.md>), [le-chonk](<https://devfeed.tech/tags/le-chonk.md>), [llamacpp](<https://devfeed.tech/tags/llamacpp.md>), [mistral](<https://devfeed.tech/tags/mistral.md>), [mistral-large-4](<https://devfeed.tech/tags/mistral-large-4.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [transformers](<https://devfeed.tech/tags/transformers.md>), [vllm](<https://devfeed.tech/tags/vllm.md>)

## AI overview

Mistral Large 4 ("Le Chonk") is an open weight multimodal model with about 1T total and 49B active parameters, a 1M token context window, and a sparse MoE decoder paired with a Pixtral vision encoder. The article emphasizes defensive cybersecurity: it reports 82% on vulnerability reproduction and patching and 93% on Cybench, while also noting strong refusal behavior on harmful cyber prompts. It describes support for Transformers, vLLM, and llama.cpp, hosted Inference Endpoints, and TRL fine tuning. Runtime and hardware needs vary by route: the vLLM NVFP4 checkpoint requires NVIDIA Blackwell GPUs, and llama.cpp's 2-bit quant can run on a single 512 GB Apple Silicon Ultra, while other listed quants require distributed inference. Reasoning effort supports low, medium, or high, defaults to high, and cannot be disabled.

## Source excerpt

We're on a journey to advance and democratize artificial intelligence through open source and open science.