CERN and Mistral sign a statement of intent for a Large Physics Model

On 15 October 2026, CERN published a statement of intent with Mistral. The page says the European laboratory and the French AI lab will work on foundation models for scientific discovery, with high-energy physics as the first focus. It is a statement of intent. The page does not describe a model that can be called, downloaded, or scored.

What they say they will train

The stated aim is one pretrained Large Physics Model. CERN calls it the first foundation model for collider physics. It would be trained on billions of simulated proton-proton collisions at the Large Hadron Collider. The tasks named on the page are reconstructing collisions, identifying particles, flagging unexpected signals, and evaluating future detector designs.

CERN describes today’s reconstruction as a chain of specialised algorithms that turn detector signals into the objects used in a physics analysis. Machine learning, it says, already improves individual steps in that chain. A foundation model, it says, could learn across the chain and cover several tasks in one model. The page does not say that chain has been replaced, and it does not give a training run that has finished.

CERN says it holds around one exabyte of experimental data from the LHC, which it glosses as one million terabytes. It says the number of collisions is set to increase tenfold with the High-Luminosity LHC, expected to begin operation around 2030. Those two figures are CERN’s account of its own data and of the upgrade.

The quotes, and what is absent

Enrica Porcari, CERN’s chief information officer, is quoted: “High-energy physics provides a unique environment in which artificial intelligence can be developed and tested with exceptional rigour. We are partnering with Mistral, a leading European AI company, to advance AI capabilities that can serve the whole physics community and be shared.”

Johannes Brandstetter, vice-president of AI for engineering at Mistral, is quoted: “We are proud to work with CERN to explore what large-scale AI models can bring to the growing complexity of particle physics. Together, we have the opportunity to push the frontiers of high-energy physics, developing AI capabilities that have the potential to lead to new scientific discoveries.”

The CERN statement gives no delivery date, no compute figure, and no budget.

What the wire adds

Reuters, carried by Channel NewsAsia, repeats the exabyte and the tenfold rise in collisions around 2030. It calls the planned model Flare, and it attributes that name to a CERN statement. The name Flare does not appear on the CERN page. Reuters also says a CERN spokesperson described the model as trained on collision data rather than text, and says no timeline or budget was disclosed. Those spokesperson lines are in the wire. The Channel NewsAsia page is labeled Reuters, and its headline is the Flare wording.

Mistral’s news index, checked on 15 October, shows no item about this statement. Mistral Large 4’s public preview is a separate release, already reported, and it is not this physics model.

What is public today is the signed intent and CERN’s description of a model still to be trained.