Google DeepMind's AI earns gold medal at International Math Olympiad
This marks a significant jump from last year, when DeepMind's earlier AI models (AlphaProof and AlphaGeometry 2) earned a silver medal with 28 points.
The IMO's President, Prof. Dr. Gregor Dolinar, confirmed the achievement, praising the AI's solutions as 'clear, precise,' and often 'easy to follow.' That's a major compliment, given that the Olympiad problems are designed to challenge some of the brightest young minds on the planet.
What makes this feat even more impressive is how the AI approached the problems. Unlike earlier systems, which needed human experts to rewrite the questions into computer-friendly formats, Gemini Deep Think worked entirely in natural language, reading and understanding the exact same problems given to human contestants, and solving them within the official 4.5-hour time limit.
The model achieved this by using advanced reasoning tools, including something DeepMind calls 'parallel thinking,' which lets the AI explore multiple problem-solving strategies at the same time, similar to how skilled mathematicians think through different angles before committing to a solution. The system was also trained using new reinforcement learning techniques and a curated set of high-quality math solutions to boost its problem-solving abilities.
Google plans to make this version of Gemini available first to trusted testers, including professional mathematicians, before rolling it out to Google AI Ultra subscribers.
The development comes amid growing competition in the AI space. OpenAI also announced that one of its experimental models hit the same score, 35 out of 42, in the same IMO challenge. With both companies now demonstrating gold-level performance on a human math competition, this signals a major leap forward in AI's ability to reason logically, handle abstract problems, and produce rigorous, formal proofs.
Google CEO Sundar Pichai described the progress as 'just astounding,' while DeepMind CEO Demis Hassabis called it 'incredible progress' in AI's ability to handle high-level intellectual tasks.
While the AI isn't replacing mathematicians anytime soon, its success opens up new possibilities for AI-assisted research, education, and mathematical discovery, and represents a milestone in machines tackling one of the most human of challenges: creative problem-solving through logic.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
an hour ago
- Yahoo
Google and OpenAI's AI models win milestone gold at global math competition
By Kenrick Cai and Jaspreet Singh (Reuters) -Alphabet's Google and OpenAI said their artificial-intelligence models won gold medals at a global mathematics competition, signaling a breakthrough in math capabilities in the race to build powerful systems that can rival human intelligence. The results marked the first time that AI systems crossed the gold-medal scoring threshold at the International Mathematical Olympiad for high-school students. Both companies' models solved five out of six problems, achieving the result using general-purpose "reasoning" models that processed mathematical concepts using natural language, in contrast to the previous approaches used by AI firms. The achievement suggests AI is less than a year away from being used by mathematicians to crack unsolved research problems at the frontier of the field, according to Junehyuk Jung, a math professor at Brown University and visiting researcher in Google's DeepMind AI unit. "I think the moment we can solve hard reasoning problems in natural language will enable the potential for collaboration between AI and mathematicians," Jung told Reuters. The same idea can apply to research quandaries in other fields such as physics, said Jung, who won an IMO gold medal as a student in 2003. Of the 630 students participating in the 66th IMO on the Sunshine Coast in Queensland, Australia, 67 contestants, or about 11%, achieved gold-medal scores. Google's DeepMind AI unit last year achieved a silver medal score using AI systems specialized for math. This year, Google used a general-purpose model called Gemini Deep Think, a version of which was previously unveiled at its annual developer conference in May. Unlike previous AI attempts that relied on formal languages and lengthy computation, Google's approach this year operated entirely in natural language and solved the problems within the official 4.5-hour time limit, the company said in a blog post. OpenAI, which has its own set of reasoning models, similarly built an experimental version for the competition, according to a post by researcher Alexander Wei on social media platform X. He noted that the company does not plan to release anything with this level of math capability for several months. This year marked the first time the competition coordinated officially with some AI developers, who have for years used prominent math competitions like IMO to test model capabilities. IMO judges certified the results of those companies, including Google, and asked them to publish results on July 28. "We respected the IMO Board's original request that all AI labs share their results only after the official results had been verified by independent experts and the students had rightly received the acclamation they deserved," Google DeepMind CEO Demis Hassabis said on X on Monday. However, OpenAI, which did not work with the IMO, self-published its results on Saturday, allowing it to be first among AI firms to claim gold-medal status. In turn, the competition on Monday allowed cooperating companies to publish results, Gregor Dolinar, president of IMO's board, told Reuters. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
New York Post
an hour ago
- New York Post
OPM keen on using AI to increase ‘efficiency' of jobs — as federal agency cuts staff by one-third
WASHINGTON — The Office of Public Management is interested in using AI to increase 'efficiency' of jobs in its quest to downsize the US government, the agency's new director told reporters Monday. Scott Kupor, a former venture capitalist, said he has a 'Silicon Valley' mentality when it comes to motivating employees and maximizing performance, and is looking to integrate AI into some of the agency's tasks. 'I view AI as probably the most significant disinflationary technology that I think we have seen yet today,' Kupor told The Post in a meeting with reporters. Advertisement 'What I mean by that is basically AI, however you deploy it, it's not gonna replace everybody's jobs, but I think AI can massively increase the efficiency of a lot of jobs, and certainly, as we know, it'll probably create new job opportunities as well.' 3 Scott Kupor, Director of the Office of Personnel Management, hopes to integrate AI into the federal agency. AP While AI won't be used to cut jobs outright, the agency has been working to drastically downsize its number of employees. Advertisement OPM is on track to cut one-third of its staff by Dec. 31, largely through voluntary departures, from 3,110 employees to around 2,000. The number of contractors used by OPM is also set to be slashed from 1,200 to about 600, the agency said. Kupor mentioned customer service and information combing — like going through the tens of thousands of comments the agency gets — as some of the tasks potentially fit for artificial technology. While the hard writing of OPM is not what he's looking to change with AI, Kupor said its 'hard to imagine' that in a few years we won't be able to 'significantly increase the efficiency' of the comment reading process 'by at least having some of that summarization, some of the kind of information gathering, information dissemination being enhanced by AI.' Advertisement 3 The Theodore Roosevelt Federal Building headquarters of the US Office of Personnel Management in Washington, DC. Getty Images As the agency's new head, he said he wants OPM employees to be 'prepared' for work in an AI 'forward way' — as President Trump has made clear he wants the US to be the global leader in AI innovation. 'I think it's a good bet that the types of jobs that we will need five or 10 years from now will be different. The types of skills we will need will be different. And I'd like to make sure that as we think about … the recruitment and talent pipeline for government,' he said. Kupor wouldn't reveal which AI companies he was most keen on using in the US government, as it's an ongoing conversation with his team. Advertisement Other government agencies have also been using AI to be more cost-efficient, especially after the Trump administration came out with policies eliminating barriers for agencies to use the technology. The Department of Veterans Affairs uses the tech to standardize care, while the Department of Justice uses it to analyze the drug market, according to a White House readout published in April. 3 Kupor said OPM will work with DOGE as much as it benefits the agency. United States Office of Personnel Management Kupor said he hopes that the OPM mindset of government efficiency transcends to other federal agencies, and that the Department of Government Efficiency was the 'catalyst' for some of the movements towards downsizing. 'To the extent that they're helpful in like the overall goals and objectives we have, then we're happy to partner with people, but we're going to ultimately make the decisions that we think are in the best interests of OPM,' Kupor said of the agency working with DOGE after Elon Musk's departure. On AI, he said he wants 'to change the mindset here' and 'we'll see if other people across the government agree where we ought to be willing to experiment with some things in a contained area with contained risk.'
Politico
an hour ago
- Politico
Trump's AI plan meets ‘energy dominance'
President Donald Trump's artificial intelligence plan — due out this week — is shaping up to be a lot like his 'energy dominance' platform. The AI Action Plan will include cutting back environmental requirements and streamlining permitting policies to make it easier to build data centers and power infrastructure, Mohar Chatterjee reports. Previous White House AI plans in the Biden administration and Trump's first term have focused on the technology's use and issues such as intellectual property. The new inclusion of permitting policy reflects both the tech industry's reality that the greatest barrier to AI deployment is access to reliable power — and Trump's enthusiasm for boosting fossil fuel production. 'Remaining the world's leader in AI will require an enormous increase in energy production,' Trump told attendees at a summit last week in Pittsburgh. The tech-energy intersection is evident in how some of AI's biggest players pitched the White House on the AI Action Plan during an open comment period this spring. Google, for example, wrote that U.S. energy and permitting rules 'appear inadequate' to meet the sector's needs. Meta called for a 'comprehensive national strategy to support domestic data center investment and deployment,' including 'timely access to reliable electricity and speed to market and permitting efficiency.' OpenAI recommended streamlining transmission line development and using the Defense Production Act to accelerate gas turbine construction. Anthropic, the company behind AI model Claude, even offered a specific energy target: 50 gigawatts of power — the equivalent of 25 Hoover Dams — dedicated to data centers by 2027. The White House has leaned on AI's needs to advance its fossil fuel agenda. The Pittsburgh summit last week was accompanied by more than $90 billion in data center and energy investments, coming on the heels of billions in announced investments during Trump's Middle East trip. A controversial report this month warning that fossil fuel plants are needed to prevent blackouts was couched in language promoting the AI race. Even as he revoked Biden-era AI and energy policies early in his term, Trump kept one that opened federal lands to data centers. He also recently launched an initiative to pair data centers with federal labs — the groundwork for a potential executive order this week. The order, which may be announced on the same day as the AI plan, is expected to direct the Energy Department to kick-start data center construction on federal sites, two people familiar with the plan told Mohar. Trump is scheduled to deliver an AI speech Wednesday at an event hosted by the All-In podcast and the Hill & Valley Forum. Green groups — which have long been concerned about data centers' potential to undermine grid reliability and increase greenhouse gas emissions — are on alert. 'Trump's AI scheme sounds like a climate bomb clothed in tech bro skin,' Jean Su, energy justice director at the Center for Biological Diversity, said Monday. It's Monday — thank you for tuning in to POLITICO's Power Switch. I'm your host, Jason Plautz. Power Switch is brought to you by the journalists behind E&E News and POLITICO Energy. Send your tips, comments, questions to jplautz@ Today in POLITICO Energy's podcast: Josh Siegel speaks with Jason Grumet, CEO of the American Clean Power Association, about the state of renewable energy under Trump. Power Centers Could geothermal get a boost from fossil fuels?Researchers and the geothermal industry are looking at ways fossil fuels can be used to generate the energy source, Shelby Webb writes. Geothermal energy has bipartisan support and the backing of federal incentives — and, crucially, the Trump administration. Trump in an executive order called for removing 'all illegitimate impediments' to geothermal projects. And Energy Secretary Chris Wright has spoken favorably about geothermal. Regulators in North Dakota approved a feasibility study in June to see whether geothermal could be paired with active oil well sites or if captured carbon dioxide could be used to feed geothermal production. Permitting gets its momentBoth Republican and Democratic lawmakers are hoping to reopen talks on permitting legislation, writes Kelsey Brugger. Committees in both the House and Senate will hold hearings this week focused on accelerating environmental reviews and addressing increased demand for electricity. 'I am hearing, believe it or not, some interest among Democrats in doing something on permitting reform. Because not only is it slowing down and grinding to a halt a lot of conventional energy projects, but there are also renewable projects, things that Democrats support,' Senate Majority Leader John Thune (R-S.D.) said during an interview Sunday on Fox News. While lawmakers in both parties want to streamline the permitting process, Democrats are wary of the GOP's intentions and the administration's unilateral actions. Need a satellite? NASA might have one for saleThe Trump administration is considering selling or leasing some NASA satellites, including ones already in orbit, Scott Waldman writes. It's in line with the administration's other efforts to slash federal spending. But NASA spokesperson Bethany Stevens stressed that no decisions have been made yet. The satellites are primarily used to gather scientific data, including carbon dioxide levels. NASA has about 50 spacecraft in orbit, and the Trump administration hasn't disclosed which ones may be up for grabs. In Other News Next stop, South America: China is building electric vehicle factories in Brazil, aiming to capitalize on one of the world's biggest markets after other automakers retreated. The little guy: There is surging interest globally in small modular reactors as nuclear energy gains steam because of soaring electricity demand. Subscriber Zone A showcase of some of our best subscriber content. The phase-out of federal renewable energy tax credits has created more uncertainty for states like New York and New Jersey that have ambitious clean energy plans. Trump dismissed all but one of the members of a federal board that oversees the storage of nuclear waste. The EPA officials who oversee the Greenhouse Gas Reduction Fund will move into the newly created Office of Finance and Administration. That's it for today, folks! Thanks for reading.
