RedNote's dots-note-3.0 Is the First AI Model to Score Perfectly at the IMO

RedNote, a Chinese social platform known in the country as Xiaohongshu, became the first company whose AI model, dots-note-3.0, scored a perfect 42 points at the International Mathematical Olympiad (IMO) in Shanghai on 20 July 2026, solving all six problems.

What happened

dots-note-3.0 solved all six IMO problems and earned 42 points, the highest possible score at the competition. According to the South China Morning Post, this is the first time in history that an artificial intelligence has achieved a full score at the IMO.

• Developer: RedNote (Xiaohongshu), a Chinese social platform
• Model: dots-note-3.0, still in beta testing
• Result: 42 out of 42 points, all six problems solved
• Place and date: the IMO wrapped up in Shanghai on 20 July 2026
• dots-note-3.0 is the lightest model in the dots3 family

What is dots-note-3.0?

dots-note-3.0 is the lightest model in RedNote's dots3 lineup, and it was still in beta when it set the record. Besides note, the dots3 family includes two larger models, jazz and aria, designed for different use cases.

RedNote (Xiaohongshu) is one of China's largest social services, known for its recommendation feed and user-generated content rather than for building foundation AI models. That makes it all the more surprising that its team was the first to take a math AI to a flawless result at the IMO.

Why a perfect score matters

42 points is the absolute maximum at the IMO: the competition consists of six problems, each worth up to seven points. Before dots-note-3.0, no AI model had gone through all six problems without dropping points, which makes RedNote's result the first perfect score by an AI in the competition's history.

The International Mathematical Olympiad has been held since 1959 and is considered the most prestigious math competition for high school students. Every year, teams from more than a hundred countries gather to solve six problems over two days. In previous years, AI models had already come close to medal-level results, but no system had managed to cleanly solve the entire set of problems until now.

IMO problems test not computational speed but the ability to build rigorous proofs and find non-obvious approaches, which used to be the territory of the strongest teenage mathematicians on the planet.

"For the first time, an artificial intelligence model has received a perfect score at the

International Mathematical Olympiad," the South China Morning Post reports.

What this means

A perfect IMO score shows that AI models have come close to the level of the best young mathematicians on proof-based and creative reasoning problems. For RedNote, a social platform rather than a dedicated AI lab, it is a bid for a place among China's technology leaders alongside other developers of frontier models. If the result is confirmed by independent verification, dots-note-3.0 will become a benchmark for assessing the mathematical abilities of future models.

Frequently asked questions

Which AI model was the first to get a perfect score at the IMO?

dots-note-3.0 from the Chinese platform RedNote. It solved all six problems and scored 42 out of 42 at the olympiad, which wrapped up in Shanghai on 20 July 2026. It is the first perfect AI result in the competition's history.

What is the dots3 family?

It is RedNote's lineup of AI models in three versions: the lightweight note (dots-note-3.0) and two larger ones, jazz and aria, built for different use cases.

What stage is dots-note-3.0 at?

At the time of the record, the model was still in beta, according to the South China Morning Post.