Alibaba’s Qwen 3 outsmarts OpenAI with perfect score in world’s toughest math contest

qwen

Alibaba’s artificial intelligence model, Qwen 3, has stunned the global tech community by achieving a perfect score in some of the world’s toughest math contests — outperforming even OpenAI’s best models. The breakthrough highlights how Chinese AI research is fast catching up to, and in some cases surpassing, Western tech giants in advanced problem-solving capabilities.

According to a report by South China Morning Post, Qwen 3 topped the leaderboard in global mathematical reasoning benchmarks, including the AIME (American Invitational Mathematics Examination) and AMC (American Mathematics Competitions) — exams typically designed to challenge the brightest human high school students. These contests test deep logic, reasoning, and multi-step problem-solving, making them an ideal way to measure how “intelligent” an AI system truly is.

What makes this milestone especially impressive is that math has long been a weakness for most large language models. While tools like OpenAI’s ChatGPT and Google’s Gemini can explain concepts or solve simple equations, they often stumble when multi-layered logic or complex calculations are involved. Qwen 3, however, seems to have cracked that barrier.

Alibaba’s researchers claim that the model not only solved problems accurately but also explained the reasoning process step by step — much like a human math prodigy would. This demonstrates that the AI’s intelligence is not just based on memorization but genuine analytical reasoning.

Experts believe this could have massive implications for how AI is used in science, finance, engineering, and research. If an AI can reliably solve complex mathematical problems, it could assist in designing new materials, optimizing algorithms, or even improving AI training itself.

This achievement also signals a growing AI rivalry between China and the US, with companies like Alibaba, Baidu, and Tencent investing heavily to compete with OpenAI, Anthropic, and Google. For now, Qwen 3’s perfect score puts Alibaba squarely on the global AI map — and perhaps even a step ahead of its American counterparts in one of the hardest tests of intelligence yet.