tech
December 15, 2025
AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms
New AI agent evolves algorithms for math and practical applications in computing by combining the creativity of large language models with automated evaluators
TL;DR
- AlphaEvolve, an evolutionary coding agent, leverages large language models (LLMs) for algorithm discovery and optimization.
- It enhances the efficiency of Google's data centers, chip design, and AI training processes.
- The agent has helped design faster matrix multiplication algorithms and found new solutions to open mathematical problems.
- AlphaEvolve uses Gemini Flash and Gemini Pro models, along with automated evaluators, to propose, verify, and score computer programs.
- Deployed algorithms have improved data center scheduling, recovering 0.7% of compute resources globally.
- It proposed a Verilog rewrite for a matrix multiplication circuit integrated into an upcoming Tensor Processing Unit.
- AlphaEvolve sped up a key kernel in Gemini's architecture by 23%, reducing training time by 1%.
- It achieved up to a 32.5% speedup for the FlashAttention kernel implementation.
- The system has rediscovered state-of-the-art solutions for 75% of tested open mathematical problems and improved solutions in 20% of cases.
- It advanced the kissing number problem by discovering a configuration of 593 outer spheres in 11 dimensions.
Continue reading
the original article