google-deepmind/alphaevolve: Repository for AlphaEvolve research and matrix multiplication discoveries

Google DeepMind has released AlphaEvolve, a repository accompanying a research paper featuring code, data, and verification notebooks for discovered results. The codebase includes a minimal local implementation, scripts and verification tools for a new upper bound on the matrix multiplication exponent, and initial programs for mathematical problems and tensor decompositions that outperform the state-of-the-art. The repository highlights that AlphaEvolve involves executing Large Language Model-generated code, which is run within a Docker sandbox for isolation. Open article →