Two Chinese mathematicians win the Fields Medal; Tsinghua invites them to teach.

In the late hours of July 23, the Fields Medal, the highest international prize in mathematics awarded every four years, was announced, with two Chinese mathematicians, Wang Hong and Deng Yu, receiving the award simultaneously.

Both were undergraduate classmates in the School of Mathematical Sciences at Peking University, class of 2007; one solved the century-old three-dimensional Kakeya conjecture, and the other resolved Hilbert's sixth problem for classical mechanics.

This is the first time mathematicians who completed their undergraduate studies in China have won the Fields Medal, and two were awarded in the same year.

Shortly after the award results were announced, Shing-Tung Yau stated at a public event at Tsinghua University: “I sincerely invite Professors Wang Hong and Deng Yu to return and join our faculty. The Tsinghua Qiu Zhen Academy is always ready to welcome them.”

This is not the first time Qiu Chengtong has competed with Peking University for talent, nor is it the first time Tsinghua University has directly challenged Peking University in mathematics.

These two of China’s most prestigious universities have been competing for decades in mathematics, the most fundamental and rigorous discipline.

From competing for teenage middle school prodigies to vying for already-renowned Fields Medalists, from university competition arenas to the new battleground of AI, from the falling out between Shing-Tung Yau and Tian Gang to today’s podium beneath the Fields Medal ceremony.

But all of this can be traced back to a departmental reorganization seventy years ago.

In 1952, during the nationwide reorganization of university departments, the entire Department of Mathematics at Tsinghua University was transferred to Peking University, including all faculty and library materials, which were moved to the Yan Garden. As a result, Tsinghua’s mathematics program was effectively discontinued for nearly 30 years, until the Department of Applied Mathematics was reestablished in 1979.

After that adjustment, Peking University’s School of Mathematical Sciences became the undisputed top choice in China’s mathematics community, with the nation’s best math talents automatically heading to Peking University.

This is where China’s “golden generation” of mathematicians emerged: Zhang Yitang, Xu Chenyang, Yun Zhiwei, and Zhang Wei—nearly all of the young mathematicians who later won the Sloan Prize, the SASTRA Ramanujan Prize, and the New Horizons Prize are graduates of Peking University.

A gold medalist in the national mathematics competition, Wei Dongyi naturally chose Peking University’s Department of Mathematics when filling out his college applications; having achieved perfect scores at two consecutive IMOs, his admission to Peking University was entirely expected.

Originally, Shing-Tung Yau wanted to help Peking University. In 1979, during his first visit to China to lecture, he went to Peking University first, where he already envisioned establishing a world-class mathematics center in China. Tian Gang was one of the students he recruited in China at that time and later brought to Harvard for his Ph.D., becoming one of his most distinguished students.

In the following two decades, he traveled back and forth, establishing mathematical institutes at the Chinese Academy of Sciences and Zhejiang University, donating books to Peking University, and giving lectures there—he even believed that the roots of Chinese mathematics would surely take hold in Yan Garden.

Around the year 2000, Shing-Tung Yau wanted to establish a platform according to international standards, opposing academic nepotism and believing that Peking University retained too many of its own students, with academic evaluations based on seniority rather than merit.

This contradiction erupted completely in 2005, causing a major upheaval in the Chinese mathematical community. Shing-Tung Yau publicly criticized Peking University’s Department of Mathematics, specifically naming Tian Gang, his own former PhD student whom he had mentored at Harvard, accusing him of academic misconduct and poor scholarly conduct, and even directly calling Peking University’s Center for Mathematical Sciences a “clique.”

A retired professor from Peking University countered: “Qiu Chengtong wanted to become the dominant figure in China’s mathematics community, but he failed.” The master and disciple completely broke ties and never spoke again. After that, Qiu Chengtong never set foot in Peking University’s School of Mathematical Sciences again.

In 2009, he directly took up the position as director of the Yau Mathematical Sciences Center at Tsinghua University. For context, since his first visit to China to lecture in 1979, over the past four decades, he has never accepted a single yuan in salary from China for his lectures or center-building efforts; instead, he personally donated $1 million to Tsinghua to support the cultivation of mathematical talent.

In 2010, he personally funded the Cheng-Tong Undergraduate Mathematics Competition, whose format directly mirrored the prestigious Putnam Competition internationally, with a difficulty level several tiers higher than typical undergraduate mathematics contests in China.

Yau's goal was clear: to create a platform under his own control that would attract the nation's best young mathematical talents to Tsinghua University. To outshine Peking University, Yau truly invested heavily.

In 2018, he established the Cheng-Tong Mathematics Talent Class, admitting no more than 30 gifted students in grades 10 and 11 each year, bypassing the Gaokao entirely for direct admission to Tsinghua University. In 2021, he went even further by founding the Qiuzhen College and launching the "Mathematics Leadership Program," recruiting students as early as ninth grade. If you demonstrate exceptional mathematical talent, you don’t need to take the Zhongkao or Gaokao—instead, you’ll undergo an eight-year integrated bachelor’s and doctoral program with top-tier faculty and fully English-language instruction, with a maximum of 100 students admitted annually.

In July 2026, 20% of the 94 students in the preparatory class were expelled for failing two core courses, Mathematical Analysis and Advanced Algebra. Over twenty parents protested at the school, and several vice presidents urged him to show leniency, arguing that such a move would make enrollment difficult next year. He dismissed them outright: “Students scoring only 20 or 30 points are just wasting resources if they stay.”

To win at Peking University, Yau Shing-Tung really went all out, showing no mercy whatsoever.

But Peking University’s Department of Mathematics has decades of accumulated strength—it’s not easy to catch up. Every year, 80% of the national mathematics Olympiad gold medals ultimately go to Peking University’s Department of Mathematics. Take Wei Dongyi, widely known as “Wei Shen,” for example—he achieved perfect scores in two IMO competitions, a feat accomplished by very few worldwide. During his undergraduate years, he won so many awards at the Qiu Cup that it became a saying in the community: “When Wei Shen competes, everyone else is fighting for second place.”

After graduation, he stayed directly at the Beijing International Center for Mathematical Research, where Tian Gang was director, and is now a living emblem of Peking University’s School of Mathematical Sciences.

The competition between the two schools has escalated from recruiting middle school students to vying for teachers, funding, and academic resources—and now, even in the hottest field of AI mathematics, both sides are pushing each other to the limit.

First, let’s talk about recruiting students. Tsinghua University’s Leading Program lowered its admission threshold to ninth grade, and Peking University quickly followed suit with its Mathematics Talent Class, targeting gifted students before their second year of high school. Every year, on the days when CMO (Chinese Mathematical Olympiad) results are announced, Tsinghua and Peking University’s admissions teams station themselves outside the exam halls, immediately approaching gold medalists and offering admission terms on the spot, afraid that a single delay might let the other side snatch them away.

Many participating students have shared that as soon as they left the exam hall, Tsinghua professors had just handed out their business cards when Peking University professors immediately pulled them into their cars.

Some students were taken aside by teachers from both sides for extended discussions at the hotel, lasting from the afternoon well into the early hours of the morning; in some cases, recruitment teams even directly promised, “If you join us, you can choose any major, and we’ll secure an additional recommendation spot for your high school.”

After recruiting students, they now go after professors. Over the years, Tsinghua University has attracted numerous young mathematicians from overseas, offering them high salaries, permanent positions, and research startup funding. Peking University has not been left behind, leveraging its decades-old alumni network to bring back many top mathematicians who previously graduated from Peking University.

In 2021, Tsinghua University recruited Caucher Birkar, a Fields Medalist aged 40, to the Yau Mathematical Sciences Center—the first time a Chinese university had directly poached a Fields Medalist still in the prime of his academic career, offering him Tsinghua’s highest-level benefits. It was as if Tsinghua had erected a golden signboard right across from Peking University.

In addition, Tsinghua University also recruited prominent young mathematicians from overseas, including Professor Liu Jun, a tenured professor in the Statistics Department at the University of Pennsylvania, and Professor Liu Zhengwei from Vanderbilt University, with Liu Zhengwei having earned his undergraduate degree from Peking University.

In 2024, Peking University invited Professor Jian Ding, a tenured professor at the Wharton School of the University of Pennsylvania, to return as a chaired professor in the School of Mathematical Sciences. Professor Ding, a 2003 undergraduate alumnus of Peking University, received the Davidson Prize, the highest award in probability theory. At the time, Tsinghua University also offered him a position, but he ultimately chose to return to Peking University.

Earlier, Peking University also brought back representatives of the "Golden Generation," such as Xu Chenyang, Yun Zhiwei, and Zhang Wei. Although Xu Chenyang later moved to MIT, Peking University’s alumni recall strategy has continued uninterrupted.

Both sides even offered nearly identical terms to secure a young scholar who had just won the Sloan Award, and in the end, it came down to who had a better personal relationship with the scholar’s advisor.

But the most intense new battleground is currently the hottest topic: AI.

Shing-Tung Yau once said: "Without mathematical breakthroughs, large models will always be mere ornaments built on parameter stacking."

In Yau’s view, the current bottleneck in large model reasoning is essentially a bottleneck in mathematical ability. Whoever can first enhance AI’s mathematical capabilities will secure entry to the next generation of large models.

Image source: Visual China

Tsinghua moved fastest, teaming up with Tencent’s HunYuan team to develop the STARE algorithm, specifically designed to solve the long-standing issue where large models get everything wrong after making a single mistake in math problem-solving. On the two most authoritative math competition benchmarks, AIME24 and AIME25, it directly improved accuracy by 4% to 8%.

Zhipu AI, developed by Professor Tang Jie’s team from Tsinghua University’s Department of Computer Science, has consistently ranked among the top tier of Chinese large models in mathematical capability, supported by a team of experts from Tsinghua’s Department of Mathematics.

At Peking University, in collaboration with the Beijing General AI Institute, we developed the TongGeometry geometric large model, which has built the world’s largest library of geometric theorems, specifically designed to solve the most challenging problems in the IMO—and even capable of discovering new geometric theorems on its own.

Previous MiniMax The M3 model that won an IMO gold medal also includes members from Peking University’s School of Mathematical Sciences on its core team.

At the 2025 WAIC World Artificial Intelligence Conference, Shing-Tung Yau personally designed the exam, bringing China’s leading large models to take a closed-book test on-site; in the end, the scores remained extremely close, with no clear winner.

Twenty years ago, China’s mathematical research didn’t rank at all on the global stage; top students graduated and headed straight for the United States, and there wasn’t even a single decent mathematical research center in the country.

However, today, one-third of the recipients on the ICCM list of distinguished Chinese mathematicians are from mainland China, triple the number from ten years ago. Tsinghua’s Qiu Zhen Institute and Peking University’s School of Mathematical Sciences now annually produce papers published in the four top mathematics journals, and several young scholars of the caliber of the Sloan Fellowship emerge each year.

Yau Shing-Tung harshly criticized Peking University back then, but he himself knew that it was precisely because Peking University was such a strong rival next door that Tsinghua University was driven to desperately build its mathematics center and intensely cultivate students.

Peking University has continuously improved its talent development system and competed for the best students precisely because Tsinghua University is closely pursuing it.

On June 28, 2026, the 17th Shing-Tung Yau College Student Mathematics Competition was held at the Meng Minwei Concert Hall at Tsinghua University, where Shing-Tung Yau himself stood on stage, facing China’s brightest group of mathematical talents seated below.

In the most prestigious individual all-around award, the Shing-Tung Yau Award, the gold medal went to Shi Haojia from Peking University, the silver medals were awarded to Lu Junpeng and Guo Huankun, also from Peking University, and the bronze medal went to Ye Yuxing, again from Peking University. The top four positions were all claimed by Peking University.

Image source: Tsinghua University official website

Six individual gold awards—the Hua Luogeng Prize, Chen Jiangsheng Prize, Zhou Weiliang Prize, Lin Jiaqiao Prize, and Xu Baolu Prize—along with the all-around award, all went to Peking University. Even the team gold award was jointly won by Tsinghua and Peking Universities.

When awarding the gold medalists from Peking University, Yau’s face wore a smile, showing no sign of anger. He even urged domestic young students in his speech to collaborate across universities to jointly advance the development of mathematics in China.

After years of struggle, Yau Shing-Tung knows clearly: what he has always wanted is not for Tsinghua University to dominate alone, but for Chinese mathematics as a whole to rise.

Back then, he traveled alone from Hong Kong to mainland China, establishing one mathematics center after another and organizing one competition after another—all for one purpose: to enable China to produce genuine mathematicians and its own Fields Medalists.

For over forty years, he never took a single yuan in salary from within the country; he paid for competitions and scholarships out of his own pocket, all for this one goal.

On July 20, an Anthropic researcher, while watching the World Cup final, had the Fable 5 model compute for several hours, then posted three simple polynomials that directly disproved the Jacobian conjecture, which had stood for 87 years and consumed the lifework of countless mathematicians.

Zhang Yitang's doctoral thesis was stuck for seven years, and it was precisely in this area.

To be fair, on this new battlefield where AI meets mathematics, Tsinghua University has taken the lead.

In China’s top tier of large models, half are from Tsinghua University. For example, Tang Jie, mentioned earlier, and his most accomplished student Yang Zhilin, who founded Moonshot AI. Kimi The latest K3 model's mathematical capabilities are comparable to those of Fable 5 and GPT-5.6 Sol, making it one of the best among domestic models for math performance.

Of course, Peking University also has people working on this. But most of Peking University’s teams focus primarily on pure academic research and haven’t produced results like Zhipu. Kimi This commercially viable, high-performance large model is indeed lagging behind Tsinghua in terms of industrial application and public awareness.

Terence Tao predicted this change three months ago in his speech at Stanford.

He says mathematics has moved from an era of "proof scarcity" to one of "proof surplus." AI can generate dozens of proofs at the level of top-tier journals in a single day, outpacing the speed at which human mathematicians can understand and verify them. He no longer tries to keep up with all new papers, as it's simply impossible to keep up.

In fact, just over a year ago, Terence Tao was criticizing AI for being too "stupid," getting eight out of ten math problems wrong.

For decades, Tsinghua and Peking University have competed over who wins and loses, but in the face of AI, that rivalry has suddenly grown much lighter. No matter who claims more gold medals or recruits more professors, everyone must now stand on the same new starting line. The era of AI-driven mathematical research has arrived—and this is the true future all mathematicians must confront.

This article is from the WeChat official account "Letter AI," authored by Miao Zheng and edited by Wang Jing.