Coding agents powered by LLMs face critical sustainability and scalability challenges in industrial deployment, with single runs consuming over 100k tokens and incurring environmental costs that may exceed optimization benefits. This paper introduces GA4GC, the first framework to systematically optimize coding agent runtime (greener agent) and code performance (greener code) trade-offs by discovering Pareto-optimal agent hyperparameters and prompt templates. Evaluation on the SWE-Perf benchmark demonstrates up to 135x hypervolume improvement, reducing agent runtime by 37.7% while improving correctness. Our findings establish temperature as the most critical hyperparameter, and provide actionable strategies to balance agent sustainability with code optimization effectiveness in industrial deployment.
@article{arxiv.2510.04135,
title = {GA4GC: Greener Agent for Greener Code via Multi-Objective Configuration Optimization},
author = {Jingzhi Gong and Yixin Bian and Luis de la Cal and Giovanni Pinna and Anisha Uteem and David Williams and Mar Zamorano and Karine Even-Mendoza and W. B. Langdon and Hector Menendez and Federica Sarro},
journal= {arXiv preprint arXiv:2510.04135},
year = {2025}
}