Salvato in:
Dettagli Bibliografici
Autori principali: Smirnova, Daria, Nasiri, Hamid, Adamska, Marta, Yu, Zhengxin, Garraghan, Peter
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2512.01099
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910137344589824
author Smirnova, Daria
Nasiri, Hamid
Adamska, Marta
Yu, Zhengxin
Garraghan, Peter
author_facet Smirnova, Daria
Nasiri, Hamid
Adamska, Marta
Yu, Zhengxin
Garraghan, Peter
contents As modern artificial intelligence (AI) systems become more advanced and capable, they can leverage a wide range of tools and models to perform complex tasks. The task of orchestrating these models is increasingly performed by Large Language Models (LLMs) that rely on qualitative descriptions of models for decision-making. However, the descriptions provided to existing LLM-based orchestrators frequently do not reflect true model capabilities and performance characteristics, leading to suboptimal model selection, reduced task accuracy, and increased cost. In this paper, we conduct an empirical analysis of LLM-based orchestration limitations and propose a cost-aware model selection method that accounts for performance-cost trade-offs by incorporating quantitative model performance characteristics within decision-making. Initial experimental results demonstrate that our proposed method increases accuracy by 0.90%-11.92% across various evaluated tasks, achieves up to a 54% energy efficiency improvement, and reduces orchestrator model selection latency from 4.51 s to 7.2 ms.
format Preprint
id arxiv_https___arxiv_org_abs_2512_01099
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Cost-Aware Model Orchestration for LLM-based Systems
Smirnova, Daria
Nasiri, Hamid
Adamska, Marta
Yu, Zhengxin
Garraghan, Peter
Artificial Intelligence
As modern artificial intelligence (AI) systems become more advanced and capable, they can leverage a wide range of tools and models to perform complex tasks. The task of orchestrating these models is increasingly performed by Large Language Models (LLMs) that rely on qualitative descriptions of models for decision-making. However, the descriptions provided to existing LLM-based orchestrators frequently do not reflect true model capabilities and performance characteristics, leading to suboptimal model selection, reduced task accuracy, and increased cost. In this paper, we conduct an empirical analysis of LLM-based orchestration limitations and propose a cost-aware model selection method that accounts for performance-cost trade-offs by incorporating quantitative model performance characteristics within decision-making. Initial experimental results demonstrate that our proposed method increases accuracy by 0.90%-11.92% across various evaluated tasks, achieves up to a 54% energy efficiency improvement, and reduces orchestrator model selection latency from 4.51 s to 7.2 ms.
title Cost-Aware Model Orchestration for LLM-based Systems
topic Artificial Intelligence
url https://arxiv.org/abs/2512.01099