CodeEvolve: an open source evolutionary coding agent for algorithmic discovery and optimization | alphaXiv
CODEEVOLVE, an open-source framework from researchers at Inter&Co., Universidade Federal de Minas Gerais, and Worcester Polytechnic Institute, integrates l
Submitted 27 Aug 2026 Inter Science - Inter&CoFederal University of Minas GeraisWPI HA Henrique Assumpção DF Diego Ferreira LC Leandro Campos FM Fabricio Murai Abstract We introduce CodeEvolve, an open-source framework that couples large language models with island-based evolutionary search for end-to-end algorithmic discovery. CodeEvolve integrates inspiration-based crossover, meta-prompting, and depth-based refinement on top of a CVT-MAP-Elites archive and a weighted LLM ensemble to generate optimized solutions for complex problems. On the AlphaEvolve benchmark suite, CodeEvolve…
saved by
related reading
- ZeroCoder: Can LLMs Improve Code Generation Without Ground-Truth Supervision?alphaxiv.org
- MirrorCode: Evidence AI can already do some weeks-long coding tasks | Epoch AIepoch.ai
- [2506.13131] AlphaEvolve: A coding agent for scientific and algorithmic discoveryarxiv.org
- Frontier Coding Agents Can Now Implement an AlphaZero Self-Play Machine Learning Pipeline For Connect Four That Performs Comparably to an External Solver — LessWronglesswrong.com
- [2501.01257] CodeElo: Benchmarking Competition-level Code Generation of LLMs with Human-comparable Elo Ratingsarxiv.org
- BigO(Bench) -- Can LLMs Generate Code with Controlled Time and Space Complexity?alphaxiv.org
- AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms — Google DeepMinddeepmind.google
- AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms — Google DeepMinddeepmind.google
- Trending Papers - Hugging Facepaperswithcode.com
- FrontierSmith: Synthesizing Open-Ended Coding Problems at Scale | alphaXivalphaxiv.org
- Mathematical exploration and discovery at scale | What's newterrytao.wordpress.com
- AlphaEvolve: How our Gemini-powered coding agent is scaling impact across fieldsdeepmind.google