What were the previous AI achievements in the IMO before this year?

Question

Accepted Answer

Previous AI Achievements in the International Mathematical Olympiad

Prior to this year's perfect scores, artificial intelligence models had demonstrated notable, though not complete, success in the International Mathematical Olympiad (IMO). In 2025, models developed by Google and OpenAI attained gold-level scores for the first time, indicating significant progress. However, even with these advancements, they were still unable to replicate the perfect 100 per cent achievement of five human contestants in that same year.
Google's Silver Medal Performance

At the 2024 IMO, Google's AI system achieved a silver-medal score. This involved successfully solving four out of the six problems presented during the multi-day competition. This performance marked a substantial step forward, showcasing the growing problem-solving capabilities of AI in complex mathematical challenges.
Full Marks Achieved by Four Leading AI Models

More recently, Deedy Das reported that four advanced AI models achieved perfect scores on this year's IMO questions. These models, developed by OpenAI, Anthropic, Axiom Math, and Moonshot AI (specifically their Kimi K3 model), each secured the maximum 42 points out of 42. This indicates a significant leap, with these AI systems matching the highest possible achievement in the contest.

Reading history

Hide →

This history is cleared when the page is closed.

What were the previous AI achievements in the IMO before this year?

Loading

Was this article helpful?

Helpful

Not helpful

Report an issue

1

0

0

Share with friends