We build reinforcement learning environments that help AI labs train models to push the boundaries of frontier math capabilities. Mathematical reasoning is a core foundation of recursive self-improvement, and we are building the infrastructure for the next major leap.
Our team combines IMO gold medalists, mathematics graduate students from Stanford and Princeton, and renowned ML researchers who understand advanced mathematics and where frontier reasoning breaks down.
While generalist RL providers work across many industries, we focus entirely on mathematics. As models advance toward research-level problems, we believe building effective environments will increasingly demand deep subject knowledge and comprehensive evaluation across multiple domains of math. We bring specialist judgment to systematically benchmark reasoning gaps in frontier models at scale, and build environments that address them.