@inproceedings{assumpcao-etal-2026-codeevolve,
title = "{C}ode{E}volve: an open-source evolutionary framework for algorithmic discovery and optimization",
author = "Assump{\c{c}}{\~a}o, Henrique and
Ferreira, Diego Leonardo Braga and
Lacerda Campos, Leandro Augusto and
Murai, Fabricio",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1345/",
doi = "10.18653/2026.findings-emnlp.1345",
pages = "28559--28581",
ISBN = "979-8-89176-462-0",
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 matches or surpasses the reported AlphaEvolve results on 5 of 9 problems and, under matched conditions, outperforms the open-source frameworks OpenEvolve and ShinkaEvolve on 6 of 9. With the open-weight Qwen3-Coder-30B backbone, it surpasses the reported AlphaEvolve score on both CirclePackingSquare instances at roughly an order of magnitude lower cost than a frontier closed-source ensemble, and remains competitive with EoH on heuristic-design tasks without retuning. Ablations show that the interaction between CodeEvolve{'}s components, rather than any single operator, drives these results. We release the framework, experimental data, and practical hyperparameter guidelines at https://github.com/inter-co/science-codeevolve."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="assumpcao-etal-2026-codeevolve">
<titleInfo>
<title>CodeEvolve: an open-source evolutionary framework for algorithmic discovery and optimization</title>
</titleInfo>
<name type="personal">
<namePart type="given">Henrique</namePart>
<namePart type="family">Assumpção</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Diego</namePart>
<namePart type="given">Leonardo</namePart>
<namePart type="given">Braga</namePart>
<namePart type="family">Ferreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leandro</namePart>
<namePart type="given">Augusto</namePart>
<namePart type="family">Lacerda Campos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fabricio</namePart>
<namePart type="family">Murai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<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 matches or surpasses the reported AlphaEvolve results on 5 of 9 problems and, under matched conditions, outperforms the open-source frameworks OpenEvolve and ShinkaEvolve on 6 of 9. With the open-weight Qwen3-Coder-30B backbone, it surpasses the reported AlphaEvolve score on both CirclePackingSquare instances at roughly an order of magnitude lower cost than a frontier closed-source ensemble, and remains competitive with EoH on heuristic-design tasks without retuning. Ablations show that the interaction between CodeEvolve’s components, rather than any single operator, drives these results. We release the framework, experimental data, and practical hyperparameter guidelines at https://github.com/inter-co/science-codeevolve.</abstract>
<identifier type="citekey">assumpcao-etal-2026-codeevolve</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.1345</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1345/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>28559</start>
<end>28581</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T CodeEvolve: an open-source evolutionary framework for algorithmic discovery and optimization
%A Assumpção, Henrique
%A Ferreira, Diego Leonardo Braga
%A Lacerda Campos, Leandro Augusto
%A Murai, Fabricio
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F assumpcao-etal-2026-codeevolve
%X 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 matches or surpasses the reported AlphaEvolve results on 5 of 9 problems and, under matched conditions, outperforms the open-source frameworks OpenEvolve and ShinkaEvolve on 6 of 9. With the open-weight Qwen3-Coder-30B backbone, it surpasses the reported AlphaEvolve score on both CirclePackingSquare instances at roughly an order of magnitude lower cost than a frontier closed-source ensemble, and remains competitive with EoH on heuristic-design tasks without retuning. Ablations show that the interaction between CodeEvolve’s components, rather than any single operator, drives these results. We release the framework, experimental data, and practical hyperparameter guidelines at https://github.com/inter-co/science-codeevolve.
%R 10.18653/2026.findings-emnlp.1345
%U https://aclanthology.org/2026.findings-emnlp.1345/
%U https://doi.org/10.18653/2026.findings-emnlp.1345
%P 28559-28581
Markdown (Informal)
[CodeEvolve: an open-source evolutionary framework for algorithmic discovery and optimization](https://aclanthology.org/2026.findings-emnlp.1345/) (Assumpção et al., Findings 2026)
ACL