AIMO-2 Winning Solution: Building State-of-the-Art Mathematical Reasoning Models with OpenMathReasoning dataset
This work addresses the problem of advancing mathematical reasoning in AI, particularly for olympiad-level tasks, by providing a comprehensive dataset and pipeline, though it is incremental in building upon existing methods.
The authors tackled the challenge of building state-of-the-art mathematical reasoning models by creating a large-scale dataset of 540K math problems with 3.2M solutions, integrating code execution through iterative training to produce 1.7M high-quality solutions, and developing a generative solution selection method, achieving top results on benchmarks.
This paper presents our winning submission to the AI Mathematical Olympiad - Progress Prize 2 (AIMO-2) competition. Our recipe for building state-of-the-art mathematical reasoning models relies on three key pillars. First, we create a large-scale dataset comprising 540K unique high-quality math problems, including olympiad-level problems, and their 3.2M long-reasoning solutions. Second, we develop a novel method to integrate code execution with long reasoning models through iterative training, generation, and quality filtering, resulting in 1.7M high-quality Tool-Integrated Reasoning solutions. Third, we create a pipeline to train models to select the most promising solution from many candidates. We show that such generative solution selection (GenSelect) can significantly improve upon majority voting baseline. Combining these ideas, we train a series of models that achieve state-of-the-art results on mathematical reasoning benchmarks. To facilitate further research, we release our code, models, and the complete OpenMathReasoning dataset under a commercially permissive license.