@inproceedings{ray-etal-2026-adaptevolve,
    title = "{A}dapt{E}volve: Improving Efficiency of Evolutionary {AI} Agents through Adaptive Model Selection",
    author = "Ray, Pretam  and
      Brahma, Pratik Prabhanjan  and
      Liu, Zicheng  and
      Barsoum, Emad",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.2019/",
    pages = "40625--40633",
    ISBN = "979-8-89176-395-1",
    abstract = "Evolutionary agentic systems intensify the trade-off between computational efficiency and reasoning capability by repeatedly invoking large language models (LLMs) during inference. This setting raises a central question: \textit{how can an agent dynamically select an LLM that is sufficiently capable for the current generation step while remaining computationally efficient?} While model cascades offer a practical mechanism for balancing this trade-off, existing routing strategies typically rely on static heuristics or external controllers and do not explicitly account for model uncertainty. We introduce \textit{ AdaptEvolve: Adaptive LLM Selection for Multi-LLM Evolutionary Refinement} within an evolutionary sequential refinement framework that leverages intrinsic generation confidence to estimate real-time solvability. Empirical results show that confidence-driven selection yields a favorable Pareto frontier, reducing total inference cost by an average of \textbf{37.9{\%}} across benchmarks while retaining \textbf{97.5{\%}} of the upper-bound accuracy of static large-model baselines."
}