AI models from Open and Google Deepmind They returned gold scores at the 2025 International Mathematics Olympiad (IMO), one of the oldest and most provocative high school mathematics competitions, companies have announced independently in recent days.
The results emphasize how quickly the AI systems are going, and yet, how uniformly the Google and Openai fit in the AI race. AI companies are strongly competing for the public perception to move on to the AI race: an intangible “Vibes” battle that can have major consequences for securing top AI talents. Many AI researchers come from background to competitive mathematics, so benchmarks such as IMO mean more than others.
Last year, Google scored a silver medal in imo Using a “formal” system, which means it required people to translate problems in a machine readable form. This year, both Openai and Google have introduced “informal” systems to the competition, which have managed to consume questions and create answers based on proof of natural language. Both companies claim that AI models responded correctly to five of the six questions about the IMO test, marking higher than most high school students and Google’s AI model from last year, without requiring any translation of a human machine.
In interviews with TechCrunch, the researchers behind the efforts of Openai and Google have claimed that these golden-medical appearances represent discoveries about AI logic models in non-verifiable areas. While AI reasoning models tend to ask questions well with simple answers, such as simple mathematical or coding tasks, these systems are struggling in tasks with more ambiguous solutions, such as buying a large chair or help with complex research.
However, Google raises questions about how Openai has conducted and announced the performance of Gold-Medal IMO. After all, if you are going to enter AI models in a mathematics competition for high students, you may also support adolescents.
Shortly after Openai’s announcement on Saturday morning, Google Deepmind Managing Director And the researchers took on social media Slam Openai for the announcement of the gold metal of premature – Shortly after IMO’s announcement, which high schools won the competition on Friday night – and for non -evaluation of their model testing by IMO.
Thang Luong, a senior researcher at Google Deepmind and head of the IMO project, told Techcrunch that Google was waiting to announce the results of IMO to respect the students who participated in the competition.
TechCrunch event
Francisco
|
27-29 October 2025
Luong said Google is working with IMO organizers from last year to prepare the test and wanted to have the blessing and official classification of the IMO President announcing its official results, which it made Monday morning.
“IMO organizers have the sorting line,” Luong said. “Thus any evaluation not based on this guideline could not submit any claim for the gold level [performance]. ”
Noam Brown, a senior Openai researcher who worked on the IMO model, told TechCrunch that IMO arrived at Openai a few months ago to participate in a formal mathematics competition, but the Chatgpt manufacturer declined because he worked on natural language systems that he thought most. Brown says Openai did not know that Imo conducted an unofficial test with Google.
Openai says it hired third-party evaluators-three former Medalists who understood the sorting system-to rate the performance of the AI model. After Openai learned of the gold-medical score, Brown said the company had arrived at IMO, which then told the company to wait to announce until the IMO awards ceremony.
IMO did not respond to TechCrunch’s request for comments.
Google is not necessarily a mistake here-passed a more formal, rigorous process of achieving the gold-medal rating-but the discussion can lose the biggest picture: AI models from many top AI workshops improve rapidly. Countries from all over the world have sent their brightest students to compete in IMO this year, and only a few percent of them scored, as well as the AI models of Openai and Google.
While Openai used a significant lead over industry, it certainly feels that the race is more careful than any company would like to admit. Openai is expected to release the GPT-5 in the coming months and the company certainly hopes to give the impression that it is still driving the AI industry.
