Mistral AI

What can large-scale AI models bring to the growing complexity of particle physics? CERN and Mistral join forces to accelerate AI for science and explore this question. We have signed a statement of intent to establish a long-term strategic partnership focused on developing foundation models for scientific discovery. CERN is home to one of the world’s largest volumes of scientific data: around one exabyte (one million terabytes) from the Large Hadron Collider, the world’s largest and most powerful particle accelerator. And with the HL-LHC, expected to begin operation around 2030, the number of collisions is set to increase tenfold. Today, analysing these collisions relies on a chain of specialised algorithms. The ambition: a single Large Physics Model, trained on billions of simulated collisions, designed to support multiple tasks: - reconstructing particle collisions, - identifying particles, - flagging unexpected signals, - evaluating future detector designs. The initial focus is on high-energy physics. But the partnership has the potential to contribute to AI approaches that can serve the broader scientific community, and lead to new scientific discoveries. European excellence in AI and fundamental science, brought together. A step towards a new generation of AI tools.

👋 Welcome to Mistral, New Material. Let’s bring design to mission-critical AI.

Mistral models are now running in orbit. 🌍. ︎ Our 3B and 8B models are onboard the first satellite of Orbitworks and Loft Orbital’s 50-satellite AI constellation, summarizing data and answering operator queries on NVIDIA Jetson AGX Orin modules. Edge AI, in the most literal sense. Read more here: https://lnkd.in/gGYA9hpD 🤝 Loft Orbital NVIDIA AI Infrastructure Orbitworks

Introducing Mistral Large 4, aka Le Chonk. Mistral Large 4 is the best open model for sovereign AI, with state-of-the-art, open-weight performance in cyber defense, manufacturing, finance and multimodal tasks, and a solid foundation for customizing models across key industries. It is a 1 trillion-parameter multimodal model with 49 billion active parameters, trained from scratch on close to 4,000 Grace Blackwell GPUs and deployed in our own data centers. Its performance is competitive with the strongest open models globally, significantly outperforming any open-weight model developed in the US or Europe, despite using a fraction of the GPU training resources. Mistral Large 4 is available to preview today with full release later this month—your feedback will help shape what ships next.

🇪🇺 The next industrial revolution will be written in Europe. Yesterday, Mistral opened its Munich AI hub: a long-term operational and research presence embedded in Europe's largest industrial economy. The hub will house Mistral's research teams dedicated to Physics and Industrial AI, and applied engineers. 🇩🇪 Why Germany? Because Germany is where the hardest engineering problems live. Automotive, energy, aerospace, advanced manufacturing: generations of proprietary process knowledge, accumulated over decades, waiting to be turned into a lasting AI advantage. 🏗️ What we are building: We're building a team of researchers, engineers, and applied AI experts who will work side by side with our customers. We're also deepening our research partnership with the Technical University of Munich. Together, we're using its wind tunnel to train models that predict aerodynamics in seconds rather than days. This marks a new chapter for Mistral in Germany, and for industrial AI in Europe. We are here to build. More here: https://lnkd.in/eiJ5ZtQS

As AI becomes critical to the next generation of space missions, we’re strengthening our collaboration with European Space Agency - ESA. Together, we will explore practical applications across Earth observation, engineering, mission design and operational decision-making, while upholding the highest standards of security, control and sovereignty. This builds on work already underway: Orbit, a secure AI assistant for ESA engineering teams, and EVE, a specialised capability for Earth observation and Earth science. More to share soon. 🚀

Today, we are announcing our partnership with Mozilla to bring privacy, control and choice to AI-powered web browsing. This represents two open source advocates working together to bring Mistral’s scientific innovations to Mozilla’s users around the world. 🌍 Together, we are building AI systems that are trained and fine-tuned on regional languages, dialects and cultural context, so anyone can get responses that understand their local nuance. Read more about the news here: https://lnkd.in/gRnECMPa

Our partnership announced today with Cloudera underscores a need for enterprises in regulated industries: having full control over their AI deployment. We believe that open, sovereign AI provides this control, and we’re excited to see what we can do together. Read more about our new partnership here: https://lnkd.in/gNTyZ2UN

Today marks a major step forward for Mistral. We are announcing a €3 billion Series D, the largest equity fundraising round ever completed by a European technology company, just three years after our launch. Led by Samsung Electronics, alongside co-leads Scaleup Europe Fund, managed by EQT Group, and PSG, the round reflects the continued support of investors and partners including ASML, NVIDIA, BNP Paribas CIB and others. Their backing is a vote of confidence in our belief that organisations need a more open, deployable and customer-controlled approach to AI. This investment will accelerate our frontier AI research, expand the infrastructure behind our models, deepen our product capabilities and support deployment at scale. We continue to believe that the future belongs to organisations that can combine sovereign state-of-the-art intelligence with the ability to deploy and adapt it in their own environments. Our open-weight models, products and infrastructure give organisations real choice over how and where they run AI, not just access to a model. What we’re seeing globally confirms it: organisations want frontier performance without becoming locked into infrastructure or governance they don’t control. That demand for sovereignty, control, choice and independence in AI, is what this round reinforces. Thank you to our customers, investors, partners and teams for backing this. Stay tuned for so much more to come! Read more here: https://lnkd.in/eYePUT9z

Mistral is bringing AI Engineer Paris back for 2026. After last year’s sold-out edition with over 700 attendees, we’re excited to announce the return of AI Engineer Paris, a two-day technical conference and expo for AI engineers, CTOs, and VPs of AI. This year, our VP of Engineering Lélio Renard Lavaud will join speakers from Black Forest Labs, Cognition, Hugging Face, and other leading AI companies to explore the future of AI engineering. The event will feature: ✦ Keynotes from industry leaders ✦ Deep technical talks and hands-on workshops ✦ An expo hall at STATION F showcasing the latest and greatest in AI See the full line-up and secure your spot: https://lnkd.in/evevXzzN