#MistralAI

Today, we release the weights of a new and improved Le Chonk, advancing the open frontier for agentic coding and cybersecurity. 💻 Built to build: Le Chonk ranks #1 on @cognition’s FrontierCode among open models, and sets a new benchmark for Pareto efficiency on @artificialanalysis. 🛡 Built to defend: State-of-the-art in cyberdefense, put through the paces by the world's leading cyber vendors, including @CrowdStrike, @cloudflare, @rubrik, @zscaler, and others. Built for sovereignty: Make it yours. Download the weights on @huggingface , customize it with Mistral Forge, and deploy it wherever you need it. Read more from our team and customers 👇:

Mistral CEO @arthurmensch on why they’re buying GPUs instead of renting: The world is supply constrained. When chips are scarce, you don’t want middlemen between you and NVIDIA. Europe’s cloud partners also couldn’t deliver at the scale they needed... so they started owning the hardware. That gives them independent token generation, the ability to train on their own compute, and faster scaling. Owning the assets raises capital needs, but he says it makes the business more sustainable if you don’t overleverage. IPO in the next few years. Full interview: 00:00 Mistral’s move beyond models 02:40 Competing with China’s AI models 05:29 Why GPU scarcity is pushing Mistral to buy, not rent 07:41 Financing the compute buildout 08:56 France, jobs and young people 10:49 AI safety and doomsday fears 12:50 Preparing for an IPO

Today, we’re introducing the Responses API in public preview, made for building agents on Mistral. Each response returns the model’s answer, its reasoning and its tool calls as separate items, so your app can handle every step more easily than with Chat Completions. It works with all chat models hosted by Mistral, including Large 4.

Congrats @MistralAI team on the next gen Le Chonk! 🙌 1T-parameter MoE multimodal model trained from scratch on 3,800 NVIDIA Grace Blackwell GPUs. Our day 0 support: NVFP4 on @huggingface → Scale distributed inference using performance-optimized Dynamo →

hey, Le Chonk by @MistralAI is now on Hub! 🐱 to celebrate, we ship day-0 support in transformers and llama.cpp, as well as quants 🤗 read our blog for model deep dive, integrations and more!

Orange Cyberdefense becomes Mistral AI's first European MSSP design Partner @orangecyberdef @orange @orangebusiness @MistralAI

Three weeks of training camp for Le Chonk. No days off. Thanks for putting it to the test, @ThibaultJaigu and @RequestyAI.

Impressed by the new @MistralAI large-4 update! The model update is much better at coding and performs as the 2nd best OSS model on our coding benchmark

@MistralAI @cognition openai : gpt anthropic : claude mistral : fat cat, french

Mistral Large 4 is now open weights, narrowly making it the #1 open weights model developed outside of China @MistralAI has updated and released the weights of their flagship model, Mistral Large 4, with 1.05T total and 52B active parameters. With high reasoning, Mistral Large 4 scores 26 on the Artificial Analysis Intelligence Index v5, up 1 point from Mistral Large 4 Preview (25), with the major differences being increased token efficiency and reduced hallucinations. It sits 1 point ahead of the leading US-developed open weights model, Muse Spark Open (xhigh, 25). On the Artificial Analysis Cyber Index, Mistral Large 4 (high) scores 51, third among current open weights models behind DeepSeek V4.1 Flash 1017 (max, 57) and MiMo-V2.6-Pro (56). Key takeaways: ➤ Narrowly leads non-Chinese open weights models: Mistral Large 4 (high) leads Muse Spark Open (xhigh), the next-best open weights model developed outside China, with the largest gap on Terminal-Bench 4.0 (25% vs. 8%). It trails leading open weights models DeepSeek V4.1 Flash 1017 (max) and GLM-5.3 (max) by 6 points on the Intelligence Index, largely due to weaker coding and agentic knowledge work performance: 15% on FrontierCode 2.0 (vs. 29% and 22%, respectively) and 1349 Elo on AA-Briefcase (vs. 1570 and 1504) ➤ Increased token efficiency: Mistral Large 4 (high) uses 82K output tokens per Intelligence Index task, 39% fewer than Preview (135K), placing it on the open weights Pareto frontier. The seven open weights models that score higher use between 113K and 225K output tokens per task ➤ #2 open weights model on the Coding Agent Index: On the Artificial Analysis Coding Agent Index, Mistral Large 4 (high) scores 48 in Mistral Vibe while using 8.5M tokens per task, about half as many as the leading open weights model GLM-5.3 (max) in Opencode (49, 15.9M). It also scores above another leading open weights model, DeepSeek V4.1 Flash 1017 (max), in DeepSeek Harness (46, 16.7M) ➤ Reduced hallucinations: Mistral Large 4 (high) has a 28% hallucination rate on the updated AA-Omniscience v1.1, down from 41% for Preview and matching Muse Spark Open (xhigh). However, its accuracy remains relatively low at 26%, similar to Ling 3.1 Flash 1019 (27%), and behind DeepSeek V4.1 Flash 1017 (max, 40%) and Muse Spark Open (xhigh, 42%) ➤ Lower price and cost per task than the Preview: Mistral Large 4 is priced at $0.70/$2.10 per 1M input/output tokens, a ~50% reduction from the Preview's pricing of $1.36/$4.18. At the new pricing, it costs $2.06 per Intelligence Index task, down 63% from $5.54 for the Preview. This is comparable to Muse Spark Open (xhigh, $2.02) but more costly than leading open weights models DeepSeek V4.1 Flash 1017 (max, $0.60) and MiMo-V2.6-Pro ($0.26) Key model details: ➤ Size: 1.05T total parameters / 52B active parameters ➤ Context window: 1M tokens ➤ Multimodality: Text and image input, text output ➤ Pricing: $0.70/$2.10 per 1M input/output tokens ($0.07 per 1M cached input tokens) ➤ License: MEOW License (Mistral Ecosystem Open-Weight License), where operating a Model-as-a-Service business requires a separate commercial license with Mistral Congratulations to the team at @MistralAI on the release!

We've just added Mistral Large 4 to Caffeine. It's Mistral's largest most capable model yet. Use it to build a new app or keep improving one you've already started. On Caffeine, its credit price is lower than Claude Sonnet 5.5's. You'll find it in the model menu, and you can switch between models in the same conversation. @MistralAI

MISTRAL 🔥: Le Chonk (Mistral Large 4) is now available on HuggingFace! > "The new model also features substantial gains in reasoning efficiency that give developers greater performance while lowering inference costs." > "Mistral Large 4 ranks as the best open-weight model for coding and among the world's most capable open models overall." It now takes the top spot on Cognition's FrontierCode benchmark among open models! We need new Mistral Small now 👀

New on Open Math OS: Mistral Large 4 by @MistralAI. 1.05T open weights. 1M context window. Available now to all approved teams.

Recent model updates from the past two weeks 👇 In Ollama’s cloud ☁️: - Beam (free for a limited time!) - Muse-spark-open - MiniMax-m3.1-flash - Mistral-large-4 - DeepSeek-v4.1-flash:1017 - Clef (Decision) Local 🦙: - Bonsai-2 - EmbeddingGemma-2 - Flint - Nemotron-3.5-bolt - Nemotron-3.5-super - Phonon (Text-to-Speech) Decision models: Clef-flash, D1, D1-omni, D3, Kev, Laya, Nimble, Nimble-v3, Pplx-decider, Strands-decider, Tev1 Try any of them after downloading Ollama: ollama run <model_name>

Mistral AI est tellement de retour !!! Très beau boulot à toutes les équipes de @MistralAI 🫡 Cocoricooooooooo 🇫🇷🇫🇷

Mistral just released a 1T model for FREE 😳 What you get: > Mistral Large 4 full weights on Hugging Face > 1M context, image input, low/med/high reasoning > Top open model on the FrontierCode coding benchmark > FREE to use and fine-tune, even for your own product How to get it: 1. Search mistralai/Mistral-Large-4-1T-A52B on Hugging Face 2. No big rig? The official demo Spaces have a FREE budget this week only 3. Or grab the Unsloth GGUFs (smallest needs ~256GB RAM) Not a laptop model, so check your hardware first.

🚨 𝗔𝗜 𝗦𝗘𝗖𝗨𝗥𝗜𝗧𝗬 𝗜𝗦 𝗠𝗢𝗩𝗜𝗡𝗚 𝗢𝗡-𝗣𝗥𝗘𝗠𝗜𝗦𝗘. Mistral has publicly released Large 4, designed to help organizations identify and fix cybersecurity vulnerabilities, with the option to run it on their own infrastructure. More control. Less dependence on external services. Would your organization trust a self-hosted AI security model? #AI #CyberSecurity #MistralAI

@MistralAI @cognition Mistral 4 as a mobile orchestrator for my tailnet + @boatdotdev microVMs Less than $1 total cost to build a fully functional boat game :D

Mistral says its new model can afford organizations a kind of control over its cyber defenses that closed models can't.

Mistral Opens Trillion-Parameter Model — Large 4 weights downloadable for cyber defense. Source: Music: Carefree, Kevin MacLeod #AI #MistralAI #cybersecurity

🤖 AI &amp; CYBERSECURITY: French AI company Mistral has publicly released its Large 4 open-weight model, designed to help organisations identify and fix cybersecurity vulnerabilities. It can be downloaded and deployed on private infrastructure. #MistralAI #ArtificialIntelligence