IMO Slams OpenAI’s Self-Proclaimed Gold Medal as 91 Judges Deny Involvement: Critics Call It “Shameless PR”

“`html

OpenAI’s claim of a new model achieving an IMO Gold medal has sparked controversy less than 24 hours after the announcement.

Multiple sources close to the International Mathematical Olympiad (IMO), including officials and academics, have voiced concerns, characterizing OpenAI’s premature declaration as “rude and inappropriate,” according to sources familiar with the matter.

The IMO organizers had requested that AI companies withhold their results until at least a week after the closing ceremony, intending to keep the focus on the young participants. OpenAI, however, seemingly jumped the gun by announcing its purported achievement shortly after the ceremony concluded.

Critics have suggested that OpenAI’s actions are consistent with a pattern of prioritizing hype over responsible disclosure, lacking both patience and official validation.

Adding fuel to the fire, questions are now being raised about the legitimacy of OpenAI’s self-proclaimed “Gold medal” performance.

Sources indicate that OpenAI was not among the AI companies officially collaborating with the IMO to test their models. Furthermore, none of the 91 official IMO judges participated in evaluating OpenAI’s submissions.

This implies that OpenAI’s “Gold medal” is a self-assessment without official endorsement, raising concerns about the rigor and objectivity of their evaluation.

What was initially perceived as a potential milestone in AI development has quickly devolved into a heated debate surrounding academic ethics and commercial promotion.

IMO officials are reportedly displeased, urging restraint and emphasizing the need to “give space to the children” participating in the competition.

The controversy stems from revelations by Joseph Myers, a veteran of the IMO since 1992, following discussions with IMO Secretary Ria van Huffel.

The consensus among IMO jury members and coordinators is that it is “rude and inappropriate” for AI developers to announce results during the IMO period, particularly before the closing ceremony. The official expectation was that AI companies would wait until at least a week after the closing ceremony before releasing any results, according to Myers.

OpenAI, represented by Noam Brown, a researcher involved in the project, acknowledged that they did not consult with IMO officials beforehand. Brown stated that they informed one organizer before publishing the results, who requested a delay until after the closing ceremony. However, this contradicts the reported request for a one-week delay following the ceremony.

What’s the real story?

A statement from Harmonic, another AI team participating in the competition, supports the “one week after the closing ceremony” timeline, specifying a deadline of July 28th.

In stark contrast to OpenAI’s assertive announcement, Google’s DeepMind has maintained a low profile. Sources suggest that DeepMind may have also achieved gold medal-level performance but has chosen to adhere to the IMO’s requested embargo, awaiting the appropriate time for disclosure.

Beyond the timing of the announcement, doubts linger about whether OpenAI’s claimed gold medal is actually justified.

Thang Luong, who leads DeepMind’s super reasoning team, added that the IMO has a specific internal grading guide not available to outsiders. Scores not complying with the guide aren’t valid for a prize.

This year’s IMO comprised six problems, each worth seven points. The gold medal threshold was 35 points. OpenAI’s self-reported score barely surpassed this mark, meaning even minor deductions during the solution process could potentially lower their score from a gold to a silver.

The IMO Gold medal saga is far from over, pending official validation from the IMO organizers.

One More Thing

Let’s turn the spotlight back to the human gold medalists.

China dominated the competition, securing the top spot with a perfect score of six gold medals and a total of 231 points. This marks their return to the world’s number one position after narrowly losing to the United States last year by a mere two points.

The United States team secured second place with five gold medals and one silver, followed by South Korea, Japan, and Poland.

The six members of the Chinese team are Deng Zhewen from Wuchang Experimental High School, Xu Qiming and Tan Hongyi from Wuhan Economic Development Foreign Language High School, Zhang Hengye from Chongqing Bashu Secondary School, Dong Zhenyu from Hangzhou Xuejun High School, and Deng Leyan from Shanghai High School.

Both Deng Zhewen and Xu Qiming previously participated in the 2024 IMO and won gold medals, with Xu Qiming being selected for the 2025 Peking University Mathematics Elite Program.

Dong Zhenyu is the only high school senior on the team who entered the national training team in his first year of high school and was recommended to Tsinghua University’s Yao Class. However, he missed being selected for the IMO national team in the previous two years.

This is his third attempt to break into the national team and finally realize his dream of completing the IMO.

Tan Hongyi was also granted admission to Tsinghua and Peking University after winning a gold medal in the China Mathematical Olympiad (CMO) for two consecutive years and being selected for the national training team.

In addition, Deng Leyan and Zhang Hengye scored a perfect 42 points in this IMO, ranking first in the world. This is also the seventh consecutive IMO since 2019 in which Chinese players have achieved perfect scores.

Deng Leyan, the youngest member of the Chinese team, only in his first year of high school, was selected for the national training team last year with a perfect score in the CMO and stood out in the IMO national team selection, setting a new record for the youngest IMO player in Shanghai in nearly 10 years.

Last year, he also won fifth place globally in the combination and probability track of the Alibaba Global Mathematics Competition, becoming the first middle school student to win a medal in the event.

Zhang Hengye won a gold medal in the 2023 CMO and was selected for the national team. He also won a prize in the Alibaba Global Mathematics Competition. After being selected for the national team last year, he took the initiative to audit classes in the Department of Mathematics at Peking University and benefited greatly.

Notably, only six of the more than 600 participants in this IMO received full marks. Almost everyone else lost in the sixth question, Combinatorial Mathematics, which was also the only question that OpenAI did not solve.

Furthermore, the next IMO will be held at Shanghai High School, which has won 18 IMO gold medals, the most in the country, and will become the first high school in the world to host the International Mathematical Olympiad.