AI systems from Google and OpenAI soar at global maths competition
To OpenAI researchers, it is another clear sign AI models can command extensive reasoning capabilities that could expand into areas beyond maths.
The optimism is shared by Google researchers, who believe AI models' capabilities 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 specialised for maths. 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 the company does not plan to release anything with this level of maths capability for several months.
This year marked the first time the competition coordinated officially with some AI developers, who have for years used prominent maths competitions such as IMO to test model capabilities. IMO judges certified the results of the 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.
OpenAI, which published its results on Saturday and first claimed gold medal status, said in an interview it had permission from an IMO board member to do so after the closing ceremony on Saturday.
The competition on Monday allowed cooperating companies to publish results, said Gregor Dolinar, president of IMO's board.
Hashtags

Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles


The South African
5 hours ago
- The South African
Why UCT has stopped policing AI-generated student work
In the global education debate on artificial intelligence (AI), the University of Cape Town (UCT) has officially adopted a university-wide framework. The roadmap – called the UCT AI in Education Framework – sets out how this tech should be integrated into teaching, learning and assessment. Central to the new policy is the decision to stop using AI detection tools, such as Turnitin's AI Score, from 1 October, citing concerns over their accuracy and fairness. According to Sukaina Walji, director of UCT's Centre for Innovation in Learning and Teaching, the framework is the result of months of consultation and development. 'AI detectors are simply not reliable…there are no magic solutions,' said Walji, while the University noted that these tools risk undermining student trust and fairness. 'We're focusing instead on assessing the process of learning, not just the product, and developing assessment strategies that are AI-resilient.' These include oral exams, collaborative projects, and reflective assignments where students must disclose how they've used AI. The Framework also includes practical support – such as training, workshops, online guides, and a six-week short course for educators – to help staff and students navigate AI responsibly. According to education professor Jonathan Jansen, South African universities are slow to adapt and what UCT is doing now should have been done five years ago. AI analyst Arthur Goldstuck echoed that view, while welcoming the move away from 'flawed' detection software. 'Many students will get AI to write the entire paper,' added Goldstuck. 'Detection tools should only be used as a basic screening mechanism, they're too unreliable to be conclusive.' 'The real danger is penalising students who've done nothing wrong.' Let us know by leaving a comment below, or send a WhatsApp to 060 011 021 1. Subscribe to The South African website's newsletters and follow us on WhatsApp, Facebook, X and Bluesky for the latest news.


eNCA
2 days ago
- eNCA
Nasdaq edges down from records ahead of big tech earnings
NEW YORK - The Nasdaq retreated from a record ahead of earnings from Google parent Alphabet and other tech giants, while European markets fretted over an August 1 deadline for the EU to avert steep tariffs from US President Donald Trump. Both the Dow and S&P 500 finished higher, but the Nasdaq dropped 0.4 percent to snap a six-day streak of records. Earnings from Alphabet and Tesla are due on Wednesday, the first two of Wall Street's "Magnificent Seven" equities to report this season. Several of the largest tech names -- such as Apple and Facebook parent Meta -- do not report results until next week. Corporate profit reports so far have painted a generally resilient picture of the US economy, but with gathering clouds in some sectors -- particularly automobiles -- from Trump's levies on major trading partners. Art Hogan of B. Riley Wealth Management said Tuesday's drop in Nvidia and some other tech names suggested profit-taking after a heady rise. "It's a difficult earnings season where expectations are really low but stocks are already priced very high," Hogan said. In Europe, only London ended the trading day in the green. Paris and Germany both finished solidly in the red. The European Union is among Washington's trading partners that face potentially steep tariffs on August 1 if they do not strike a deal with Trump's administration. "European markets have been getting increasingly jittery as the (August 1) deadline approaches," said David Morrison, senior market analyst at Trade Nation. "With little sign of progress so far, investors are preparing for possible tariff retaliation from the EU." US Treasury Secretary Scott Bessent said he would meet his Chinese counterparts in Stockholm next week for tariff talks, as a separate mid-August deadline approaches for US levies on China to snap back to steeper levels.


Mail & Guardian
2 days ago
- Mail & Guardian
AI can advance the sustainable development goals
Artificial intelligence used with intention, inclusivity and oversight improve healthcare and education, as well as mitigate the effects of climate change. Photo: Delwyn Verasamy/M&G As South Africa contends with persistent development issues such as poverty, inequality, healthcare disparities, educational gaps and environmental threats, the promise of artificial intelligence (AI) is no longer a distant frontier but an essential catalyst for transformative change. AI is already being implemented globally to tackle complex development problems. For South Africa, the question is how to integrate it effectively to support sustainable and inclusive growth. If wisely and ethically harnessed, AI could become one of the most powerful instruments in South Africa's pursuit of the United Nations' sustainable development goals (SDGs). It has the potential to accelerate progress, narrow inequality and unlock the latent potential of South African society. But this potential must be cultivated with intention, inclusivity and oversight. SDG 3: Transforming healthcare and saving lives: South Africa's healthcare system, although marked by moments of excellence, remains strained by disparities in access, quality and resource allocation. The application of AI in this domain holds promise not just for efficiency but for equity. AI-driven diagnostic tools can rapidly detect diseases such as tuberculosis, malaria and cancer. Algorithms trained on large datasets can identify symptoms from images or scans with an accuracy that rivals, or even exceeds, that of human practitioners. Moreover, AI can improve disease surveillance by predicting and modeling outbreaks, which is critical in a country still grappling with the dual burden of communicable and non-communicable diseases. Predictive models based on environmental data, patient mobility and historical records can help public health authorities anticipate and mitigate disease spread before it escalates into full-blown crises. Treatment personalisation is another frontier. AI systems can optimise treatment plans based on a patient's genetic profile, lifestyle data and real-time biometrics, thereby enhancing outcomes and reducing adverse effects. This is particularly relevant for chronic disease management such as diabetes, hypertension, and HIV/Aids, where continuous monitoring and dynamic response are key. In remote or underserved areas, AI-powered mobile health platforms can bring diagnostic and consultation services to communities long excluded from specialist care. SDG 4: Education: In the realm of education, AI is poised to democratise access and personalise learning experiences in a manner previously unimaginable. South Africa's education system, despite significant investment, remains beset by inequality in quality and outcomes, especially in rural and peri-urban areas. AI can bridge these gaps through intelligent tutoring systems that adapt to individual learners' pace, preferences and proficiency. For instance, AI-powered platforms can assess where a student is struggling and adjust content delivery to reinforce those areas, offering real-time feedback and customised learning paths. This degree of personalisation can significantly reduce dropout rates and improve performance across diverse learner populations. Furthermore, AI can support inclusive education by assisting learners with disabilities. Speech-to-text, text-to-speech and natural language processing tools can make content more accessible to blind, deaf or dyslexic students. In multilingual societies like South Africa, AI-driven language translation tools can also ensure that learning materials reach students in their home languages, enhancing comprehension and cultural relevance. Educators also benefit. AI can automate administrative tasks, provide insights into student performance and suggest interventions, freeing teachers to focus on pedagogy and mentorship. At a systemic level, AI can support policymakers by analysing educational outcomes across regions and demographics, enabling targeted interventions and better allocation of resources. SDG 13: Climate change: As the climate crisis intensifies, South Africa finds itself on the front line of its economic, social and ecological consequences. Water scarcity, biodiversity loss and extreme weather events pose existential threats to both urban and rural livelihoods. AI offers sophisticated tools for mitigation, adaptation and resilience-building. Through the processing of vast environmental datasets, AI can produce accurate climate models that forecast the effect of rising temperatures, shifting rainfall patterns and other ecological disruptions. AI can integrate real-time weather data, satellite imagery and hydrological models to improve drought forecasts. These models can guide agricultural planning, infrastructure development, and water resource management with unparalleled precision. Farmers can use AI for precision agriculture, monitoring soil health, optimising irrigation and predicting yields. Energy efficiency is another critical area. AI can optimise the generation, distribution and consumption of energy, reducing carbon footprints while improving access. Smart grids informed by machine learning algorithms can predict demand spikes and reroute electricity to prevent outages. During disasters such as floods or wildfires, AI-enabled systems can provide early warnings, simulate response strategies and coordinate relief efforts. Satellite data combined with AI analytics can map affected areas in real time, identify vulnerable populations and facilitate targeted humanitarian interventions. These applications are not just technically sophisticated but socially vital. Integrating ethics and equity into AI development But the journey to AI-enabled SDGs is not without risks. Issues of data privacy, algorithmic bias, surveillance and digital exclusion must be confronted directly. The use of AI must be guided by principles of transparency, accountability and justice. Local contexts matter, and solutions must be co-created with communities, grounded in local knowledge and aligned with national development priorities. Moreover, AI must not deepen inequalities by creating a technological elite. Investments must be made in digital infrastructure, skills development and research capabilities, particularly in historically marginalised areas. If AI is to be a force for good, its benefits must be broadly shared, and its design must reflect the values and diversity of South African society. To avoid surveillance, bias and misuse of data, the country needs strong ethical guidelines. The Presidential Commission on the Fourth Industrial Revolution has made initial recommendations, but these must be translated into enforceable policies. The private sector, particularly in fintech, agri-tech, edtech and healthtech, must be encouraged to innovate responsibly. Universities and research institutions should intensify efforts to localise AI knowledge production and ensure that South African problems are being solved by South African minds. Most importantly, the state must play a catalytic role in ensuring that the regulatory frameworks, data governance standards and public investments align with the broader vision of sustainable development. The alignment between AI and the SDGs is not coincidental; it is foundational. AI is not just about machines, it is about leveraging intelligence, in all its forms, to solve humanity's greatest problems. Let us seize this opportunity not just with code and computation, but with compassion, conscience and collective purpose. Dinko Herman Boikanyo is an associate professor of business management at the University of Johannesburg. He writes in his personal capacity.