Advancing physical reasoning in open models.
AIPhO challenges the open-source community to build open-weight models that can reason through hard, original physics problems, with a prize pool of over $250,000.
Our mission
Solving a physics problem takes more than recalling formulas. It means turning a real situation into a model, choosing what to neglect, and checking whether an answer makes physical sense. These are skills AI systems still struggle with.
The most capable reasoning models today are closed. Researchers can use them, but they can’t inspect them, adapt them, or build on how they work. For science, that matters: a result you can’t reproduce is hard to trust.
AIPhO exists to change that for physics. Every competing model is open-weight, and winning approaches are shared with the community, so progress in physical reasoning belongs to everyone, not just the few organizations that can afford to train frontier models.
As of October 2026, even the best publicly available model, GPT-5.6 Sol, solves just 32% of CritPt, a benchmark of research-level physics written by more than 50 working physicists. The strongest open models trail close behind, at about a quarter. Open models are within reach of the frontier, and AIPhO is built to close the gap.
The competition
Teams train and fine-tune open-weight models, then submit them through Kaggle to be evaluated on a hidden set of 110 original problems, written from scratch so they appear in no training data.
Problems range from international olympiad style to the level of active research, and are written by IPhO and EuPhO gold medalists alongside physicists from MIT, Oxford, the University of Toronto and other universities. The competition is open to anyone: companies, university labs, independent researchers and students.
Competition dates and full rules will be announced soon.
Who we are
AIPhO’s team includes the organizers of the Online Physics Olympiad, which has brought together more than 10,000 competitors from over 80 countries across six years. They’re joined by active researchers, PhD students, and international olympiad medalists, coaches and problem writers.
AIPhO is an independent project modeled on the AI Mathematical Olympiad, whose team advises us.
Get involved
Sponsor the competition
We're growing the prize pool to bring in more researchers and strengthen the open-source community.
Write problems
We're hiring physicists who can write original problems that current state-of-the-art models cannot solve.
Compete
Registration opens on Kaggle when the competition launches.