Newest papers

Which Model for Which Role? Cost-Efficient Pareto Recommendation for LLM Agentic Workflows

Qian Xie, Wenyue Hua, Yueli He, Sripad Karne, Armaan Agrawal, Nikos Pagonas, Eugene Wu, Kostis Kaffes, Tianyi Peng, Nairen Cao · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26676 · cited by 0 works (OpenAlex) · open-access PDF

German

LLM Agent Workflows verwenden zunehmend verschiedene Modelle für spezialisierte Rollen. Benutzer können verschiedene Kompromisse zwischen Aufgabenqualität und Bereitstellungskosten bevorzugen, manchmal auch unter Berücksichtigung der Latenz. Gute Aufgaben finden eine separate Suchkosten aus der Bewertung mehrerer Kandidaten. Wir stellen Kosten gekoppelte Gittins (CC-Gittins) ein, eine multi-objektive Erweiterung der ein-Ziel-kostenbewussten Gittins-Suche nach Pareto-Empfehlung. Es nutzt die Verbindung zwischen Bewertung und Bereitstellungskosten: Aufgaben, die mehr kosten, um zu implementieren, kostet auch mehr, um zu bewerten. Die Suche alterniert zwischen der Suche nach höherer Qualität und niedrigeren Bereitstellungskosten, wobei jede Bewertung wieder verwendet wird, um beide Such

Machine translation (Meta M2M100); the English original follows.

Abstract

LLM agentic workflows increasingly employ different models for specialized roles. Users may prefer different trade-offs between task quality and deployment cost, sometimes also considering latency. Finding good assignments incurs a separate search cost from evaluating multiple candidates. We introduce Cost-Coupled Gittins (CC-Gittins), a multi-objective extension of single-objective cost-aware Gittins search for Pareto recommendation. It exploits the coupling between evaluation and deployment costs: assignments that cost more to deploy also cost more to evaluate. The search alternates between seeking higher quality and lower deployment cost, reusing each evaluation to inform both searches. After each batch, CC-Gittins recommends assignments that are nondominated under uncertainty-adjusted estimates of quality and deployment cost. We evaluate CC-Gittins on eight quality--cost matrices from workflows with two to five roles and up to 1,681 candidate assignments. We also evaluate the method with latency as a third objective. Across the eight benchmarks, CC-Gittins recovers a substantial fraction of the true Pareto frontier at roughly 10% of the exhaustive-evaluation cost, and recovers most of the frontier with few false positives by 30%.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Workflow, Software deployment, Computer science, Pareto principle, Exploit, Task (project management)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange