AI Models””https://x.com/alexwei_/status/1946477742855532918″ ” rel=””noreferrer noopener nofollow”” target=””_blank” “> Openai Google DeepMind has achieved gold medal scores in the 2025 International Math Olympiad, one of the oldest and most challenging high-school-level math competitions in the world. The companies announced this independently in recent days.
These results show how quickly AI systems are improving, but also how evenly Google and OpenAI appear to be matched in the AI race. AI companies are fiercely competing for the public perception that they are ahead in the AI race. This is an intangible “vibe” battle that can have a big impact on securing top AI talents. Many AI researchers have backgrounds in competitive mathematics, so benchmarks such as IMO are more important than others.
The year before last Google won a silver medal at IMO using a “formal system”which required humans to translate the problems into a machine readable format. Both OpenAI and Google entered their “informal systems” into the competition this year. These systems were able ingest questions and produce proof-based responses in natural language. Both companies claim that their AI models answered five out six questions correctly on IMO’s tests, scoring higher than the majority of high school students, and Google’s AI from last year, all without requiring any translation between human and machine.
In interviews given to TechCrunch by researchers from OpenAI and Google, they claimed that the gold-medal performances represented breakthroughs in AI reasoning models for non-verifiable areas. AI reasoning models are good at simple math and coding, but struggle with tasks that require more ambiguous answers. For example, buying a chair or helping to conduct complex research. Google, however, is raising concerns about how OpenAI announced and conducted its gold-medal IMO results. If you’re going enter AI models in a math competition for high schoolers you might as well argue with teenagers.
On Saturday morning, OpenAI announced their feat. Google DeepMind CEO [and researchers]took to social media Slam OpenAI for announcing their gold medal prematurely – shortly after IMO announced the high schoolers who had won the competition Friday night – and for not having IMO officially evaluate their model’s tests.
Btw, as an aside: we didn’t make the announcement on Friday because we respected IMO Board’s request that all AI Labs share their results after the official results have been verified by independent experts & students have received the acclamation and recognition they deserve
– Demis Hassabis””https://twitter.com/demishassabis/status/1947337618787615175?ref_src=twsrc%5Etfw”” rel=””nofollow” “>
Techcrunch event
San FranciscoOctober 27-29, 2025
Luong said that Google has been working with IMO’s organizers since last year in preparation for the test and wanted to have the IMO president’s blessing and official grading before It announced its official results on Monday morning.
Luong said that the IMO organizers had their own grading guidelines. “Any evaluation that is not based on this guideline cannot make any claims about gold-medal levels [performance].” Brown claims that OpenAI was unaware that IMO was testing with Google in an informal manner.
OpenAI claims it hired third-party evaluators – three former IMO Medalists who understand the grading systems – to grade its AI model’s performance. Brown said that after OpenAI learned its gold-medal result, the company contacted IMO. IMO then told the company to delay announcing the news until after IMO’s Friday night award presentation. TechCrunch did not receive a response from IMO regarding its request for comment.
Google’s not necessarily wrong — it did undergo a more formal, rigorous process in order to achieve its gold medal score — but this debate may miss the larger picture: AI models are improving rapidly from several leading AI laboratories. Just a few percent scored as well as OpenAI’s and Google’s AI models at IMO. This year, countries from all over the world sent their brightest student to compete.
OpenAI used to be a leader in the AI industry. However, it seems that the race is closer than any company would admit. OpenAI will release GPT-5 within the next few months. The company hopes to maintain its position as the leader in the AI industry. Maxwell Zeff, a senior reporter for TechCrunch who specializes in AI, is available at
. Zeff covered the rise and fall of AI, as well as the Silicon Valley Bank Crisis, for Gizmodo and MSNBC. He is based out of San Francisco. When he is not reporting, you can find him hiking, biking and exploring the Bay Area food scene.
View Bio
