# CERN and Mistral Plan AI Model for Particle Physics Research By Simon Yoon Canonical URL: https://www.tokenpost.com/news/technology/32510 Published: 2026-10-15T12:46:05.000Z Updated: 2026-10-15T12:46:05.000Z Section: Technology > The Large Physics Model would train on billions of simulated LHC collisions and support reconstruction, particle identification, signal detection and detector design. CERN and Mistral AI signed a statement of intent Thursday, Oct. 15, to develop a particle physics model aimed at handling analysis needs as collider data volumes rise. The initiative, called the Large Physics Model, is the first project in a long-term scientific AI partnership. It would be trained on billions of proton-proton collision events generated by simulation at the Large Hadron Collider (LHC). Planned uses include rebuilding collision events, identifying particles, flagging unexpected signals and helping researchers evaluate possible detector designs. The proposal covers several parts of the research workflow rather than a single narrowly defined task. Today, collision reconstruction uses specialized algorithms to turn detector readings into particles and other objects used in research. Machine learning already supports some focused tasks. The new model is intended to cover more of that analysis chain in one system. No performance results have been presented to show how well it can handle the planned work. The training material and stored archive are different: planned training would use simulated events, while the experimental-data tally reflects records already collected. CERN has stored about one exabyte of LHC experimental data, roughly 1 million terabytes. No additional details on the planned training set have been announced. The High-Luminosity Large Hadron Collider is expected to begin operating around 2030. The upgraded machine is designed to produce roughly 10 times as many collisions and as much data as the original LHC design. That projected growth would increase the volume of material researchers need to analyze. “High-energy physics provides a unique environment in which artificial intelligence can be developed and tested with exceptional rigour,” said Enrica Porcari, CERN’s chief information officer. “We are proud to work with CERN to explore what large-scale AI models can bring to the growing complexity of particle physics,” said Johannes Brandstetter, Mistral’s vice president of AI for Engineering. A project schedule, funding figure, computing resources, milestones, access terms and performance benchmarks remain unspecified. CERN has not identified [Mistral Large 4]() as part of the physics project, and the partners have not announced when development or training will begin. ## Links in this article - [Mistral Large 4](https://www.tokenpost.com/news/technology/27063)