LONDON, April 2, 2024 -- XTX Markets is excited to announce the opening of the first $1.048 million progress prize for the Artificial Intelligence Mathematical Olympiad (AIMO), on the Kaggle platform.

This initiative is part of the $10 million AIMO Prize, launched in November 2023, aimed at fostering the development of AI models capable of mathematical reasoning. The ultimate goal is to create a publicly shared AI model that could achieve a gold medal standard at the International Mathematical Olympiad (IMO).

The grand prize of $5 million will be awarded to the first AI model that competes in an AIMO Prize-approved competition and attains a performance equivalent to an IMO gold medal.

To encourage continuous progress towards this goal, XTX Markets has introduced a series of progress prizes.

This first progress prize includes problems at intermediate-level high school mathematics competitions, but do not reach IMO difficulty. It is open from 1 April to 27 June, with the prizes to be distributed to up to five teams in July 2024 at the 65th IMO in Bath, UK.

First Progress Prize Details:

  • $1.048 million prize pool.
  • The Kaggle platform's private leaderboard will determine the distribution of $253,952 among the top five teams, based on minimum performance criteria.
  • Remainder of the pool will go to the first team scoring at least 47/50 on both public and private leaderboards. If not won, it will roll over to the next year's prize pool.
  • Eligibility for prizes requires the open publication of the teams' code, methodology, data, and model parameters.

For more competition details, visit the Progress Prize participant page on Kaggle: https://www.kaggle.com/competitions/ai-mathematical-olympiad-prize/, the AIMO Prize website: https://aimoprize.com/ or join the conversation on Reddit: https://www.reddit.com/r/AIMOprize/.

Dr Geoff Smith MBE, Honorary Reader in Mathematics at the University of Bath, commented:

"I am delighted that XTX Markets have launched a serious, objective test for AI. It is all very well to simulate casual human blather, but hopefully this will see if it can begin to simulate proper reasoning."