OpenAI’s Controversial IMO Gold Claim Sparks Ethics Debate

OpenAI claimed that its new model achieved a gold‑medal score at the International Mathematical Olympiad (IMO) less than 24 hours after the competition, triggering a dramatic reversal in the narrative.

Multiple IMO officials and leading academics described OpenAI’s actions as “rude and inappropriate,” arguing that the organization should have waited at least a week after the closing ceremony to announce results, thereby keeping the focus on the teenage participants.

The controversy deepened when it emerged that OpenAI was not one of the AI companies officially collaborating with the IMO; none of the 91 IMO judges evaluated its answers, meaning the claimed gold‑medal result lacked official verification.

IMO Officials Angry: “Give Space to the Kids”

IMO veteran Joseph Myers spoke with IMO Secretary‑General Ria van Huffel, emphasizing that AI developers should wait at least a week after the closing ceremony before releasing results, to keep attention on the young competitors.

OpenAI’s representative Noam Brown admitted that the company had not contacted IMO officials in advance and only informed a single organizer, who requested that the results be announced after the closing ceremony—contradicting the public demand for a one‑week delay.

Another AI team, Harmonic, confirmed the “one‑week after closing ceremony” requirement, specifying a deadline of July 28.

In contrast, Google DeepMind, while reportedly achieving gold‑medal‑level performance, chose to respect the IMO’s timing request and remained silent.

Beyond the timing dispute, the validity of OpenAI’s gold‑medal claim is also questioned. IMO internal scoring guidelines are not publicly available, and without them, any external score lacks official standing.

The IMO this year featured six problems worth seven points each, with a gold‑medal threshold of 35 points. OpenAI’s self‑reported score barely crossed this line, meaning even minor deductions could drop it to silver.

The final official results placed the Chinese team at the top with six gold medals and a total of 231 points, followed by the United States, South Korea, Japan, and Poland.

Individual Chinese contestants were highlighted, including Deng Zhewen, Xu Qiming, Tan Hongyi, Zhang Hengye, Dong Zhenyu, and Deng Leyan, many of whom have previously earned full marks or top honors in national and international math competitions.

The article notes that only six out of over 600 participants achieved perfect scores, all excelling in combinatorial mathematics—the one problem OpenAI failed to solve.

The next IMO will be hosted by Shanghai High School, which has accumulated 18 gold medals to date, making it the most successful Chinese school in IMO history.

Observers suggest that future mathematics competitions may feature a three‑way rivalry among China, the United States, and artificial intelligence.