Discovering Algorithms with Reinforcement Learning

We train frontier models to discover novel solutions for the most important problems.