AI Excels at the International Mathematical Olympiad for the First Time
![]() |
| AI and Mathematics |
For the first time, artificial intelligence has achieved a perfect score in the world's premier mathematics competition—a feat previously limited to a select group of gifted young people. Chinese technology companies Huawei and Xiaohongshu announced that their respective AI models both achieved a perfect score (100%) on the International Mathematical Olympiad (IMO) questions that human competitors faced this month.
Huawei and Xiaohongshu were the first AI companies to release their results, and other companies worldwide are expected to follow suit in the coming days. In a statement this week, Xiaohongshu said, "We are delighted with this result, as achieving a perfect score in the International Mathematical Olympiad is a significant challenge."
![]() |
| AI and Mathematics |
AI and Mathematics
The company added, "No large language model had ever achieved a perfect score under the official International Mathematical Olympiad evaluation criteria"—referring to the technology underpinning AI tools like chatbots. In 2025, models from companies such as Google and OpenAI achieved gold-medal-level results for the first time, yet they could not match the performance of the five human contestants who secured perfect scores.
This year, 666 contestants from various countries participated in the International Mathematical Olympiad (IMO) held in Shanghai. According to official results, only seven of them achieved a perfect score. Participants in the IMO—which is hosted in a different location each year—must be under the age of 20.
On Wednesday, Huawei announced that its AI-powered program, "Celia," demonstrated "comprehensive problem-solving capabilities" across several mathematical fields. Meanwhile, Xiaohongshu—known in English as "Red Note"—announced that its model, "Dots Note 3.0," participated in the Olympiad for the first time this year.
Tech companies did not receive the IMO questions until human contestants had finished the test, and they were required to submit their answers within a specific timeframe. Xiaohongshu emphasized that "any form of human intervention during the test was strictly prohibited," noting that the answers provided by the AI tools were submitted to the Olympiad organizers for evaluation.
In the 2024 Math Olympiad, Google won a silver medal, successfully solving four out of the six problems over a period of two to three days. Deedy Das, a partner at the US investment firm Menlo Ventures, stated that he was the one who presented this year's Math Olympiad questions to four advanced AI models.
He added on LinkedIn that the four models—from OpenAI, Anthropic, Axiom Math, and Moonshot AI—all achieved a perfect score of 42 out of 42. "AI has officially surpassed the boundaries of mathematics in the International Mathematical Olympiad," Das wrote.


Leave a Comment