For the first time in history, an AI model has earned a perfect score at the prestigious International Mathematical Olympiad (IMO). Developed by Chinese tech giant RedNote, the model named dots-note-3.0 solved all six competition problems flawlessly, marking a breakthrough in artificial intelligence’s problem-solving capabilities.
- Dots-note-3.0 achieved a perfect score of 42 at the IMO
- AI processed original problems without human translation or intervention
- RedNote plans to open-source the model in the near future
What happened
In the 2026 International Mathematical Olympiad held in Shanghai, RedNote’s AI model dots-note-3.0 made history as the first ever artificial intelligence to earn a perfect score. The model solved all six challenging mathematical problems, each requiring rigorous proof and logical argumentation rather than just final answers. The IMO is a high-level global competition established in 1959 and known for its difficulty, making this achievement a significant milestone.
Unlike earlier AI contenders that relied on human experts to convert problem statements into formal logic or programming languages, dots-note-3.0 directly processed the original competition documents independently. It operated under the official contest rules prohibiting any human help, including hints or corrections, showcasing a wholly autonomous reasoning system.
Why it matters
The perfect score by dots-note-3.0 represents an unprecedented advancement in AI’s capability to understand and solve complex mathematical reasoning tasks end-to-end. This contrasts with prior AI performances such as Google DeepMind and OpenAI models, which achieved high but imperfect scores and required human assistance for problem translation or reasoning verification. RedNote’s approach integrates natural language thinking with programmatic proof testing and refinement.
This breakthrough has important implications for the future of AI-driven research and education, as it demonstrates that AI can independently tackle abstract, logical problems traditionally considered the domain of human mathematicians. It also establishes a new benchmark for AI applications in automated theorem proving, critical thinking, and scientific discovery.
What to watch next
RedNote has announced intentions to open-source dots-note-3.0 soon, which could catalyze wider adoption and development of advanced reasoning AI in academia and industry. Observers will watch closely to see how accessible the model becomes and whether it can be adapted for broader real-world mathematical or technical problem-solving beyond competition settings.
Additionally, the progress of other AI systems from major tech research groups in upcoming IMO events or similar contests will be monitored to evaluate if this breakthrough sparks a new wave of competitive innovation in AI reasoning capabilities. The evolving interplay between human and AI performance in mathematics will shape future directions for both education and automated discovery tools.