Xiaohongshu's dots-note-3.0 Scores Perfect 42/42 at IMO 2026

Xiaohongshu's internal model, dots-note-3.0, was invited to participate in the official evaluation of the 67th International Mathematical Olympiad (IMO 2026) and won a gold medal with a perfect score of 42/42. According to the official Xiaohongshu Tech REDtech account and media reports, this marks the first time a Chinese large language model has achieved a perfect score in the official IMO evaluation, making it the second model globally to accomplish this feat. The model's mathematical reasoning capabilities are now on par with top human contestants, as only 7 human participants achieved a perfect score in this year's competition.

Confirmed

Official information confirms that dots-note-3.0 correctly answered all six problems. Technologically, the model demonstrated robust end-to-end proving capabilities, distinguishing itself from traditional approaches that rely on human translation of problems into formal languages.

Unconfirmed

Tech blogger @op7418 mentioned that this lightweight model will be open-sourced in the future, which would provide new resources for the tech community, though a specific timeline or scope for the open-source release has not yet been detailed.

2026-07-21 ~ 2026-07-23 · 7 related posts

Primary sources