Team of Thoughts: Efficient Test-time Scaling of Agentic Systems through Orchestrated Tool Calling

Jeffrey T. H. Wong, Zixi Zhang, Junyi Liu, Yiren Zhao

Keywords: Multi-Agent Systems, Large Language Model

TL;DR: We propose Team of Thoughts, enabling heterogeneous agents to collaborate efficiently through dynamic orchestrator selection and parallel tool agents execution, achieving superior task performance and cost efficiency.

Abstract: Existing Multi-Agent Systems (MAS) typically rely on static, homogeneous model configurations, limiting their ability to exploit the distinct strengths of differently post-trained models. To address this, we introduce \textbf{Team-of-Thoughts}, a novel MAS architecture that leverages the complementary capabilities of heterogeneous agents via an orchestrator-tool paradigm. Our framework introduces two key mechanisms to optimize performance: (1) an orchestrator calibration scheme that identifies models with superior coordination capabilities, and (2) a self-assessment protocol where tool agents profile their own domain expertise to account for variations in post-training skills. During inference, the orchestrator dynamically activates the most suitable tool agents based on these proficiency profiles. Experiments on five reasoning and code generation benchmarks show that Team-of-Thoughts delivers consistently superior task performance. Notably, on AIME24 and LiveCodeBench, our approach achieves accuracies of 96.67\% and 72.53\%, respectively, substantially outperforming homogeneous role-play baselines, which score 80\% and 65.93\%. We open-source our code at here.

Submission Number: 69