![[2510.11608] ParaCook: On Time-Efficient Planning for Multi-Agent Systems](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2510.11608] ParaCook: On Time-Efficient Planning for Multi-Agent Systems
Summary
ParaCook introduces a benchmark for time-efficient planning in multi-agent systems, focusing on collaborative tasks inspired by cooking games, revealing limitations in current LLM approaches.
Why It Matters
This research addresses a critical gap in existing benchmarks for multi-agent systems by emphasizing time efficiency in planning. As AI systems become more integrated into real-world applications, understanding their collaborative capabilities and limitations is essential for advancing AI technology and improving operational efficiency.
Key Takeaways
- ParaCook benchmarks time-efficient planning in multi-agent systems.
- Current LLMs struggle with parallel actions and coordination.
- The benchmark is inspired by the Overcooked game, simulating real-world tasks.
- LLMs show potential in high-level parallel optimization tasks.
- The framework allows for scalable evaluation with adjustable complexity.
Computer Science > Artificial Intelligence arXiv:2510.11608 (cs) [Submitted on 13 Oct 2025 (v1), last revised 15 Feb 2026 (this version, v2)] Title:ParaCook: On Time-Efficient Planning for Multi-Agent Systems Authors:Shiqi Zhang, Xinbei Ma, Yunqing Xu, Zouying Cao, Pengrui Lu, Haobo Yuan, Tiancheng Shen, Zhuosheng Zhang, Hai Zhao, Ming-Hsuan Yang View a PDF of the paper titled ParaCook: On Time-Efficient Planning for Multi-Agent Systems, by Shiqi Zhang and 9 other authors View PDF HTML (experimental) Abstract:Large Language Models (LLMs) exhibit strong reasoning abilities for planning long-horizon, real-world tasks, yet existing agent benchmarks focus on task completion while neglecting time efficiency in parallel and asynchronous operations. To address this, we present ParaCook, a benchmark for time-efficient collaborative planning. Inspired by the Overcooked game, ParaCook provides an environment for various challenging interaction planning of multi-agent systems that are instantiated as cooking tasks, with a simplified action space to isolate the core challenge of strategic parallel planning. Through a comprehensive evaluation of state-of-the-art LLMs, we find that current approaches achieve suboptimal plans, which struggle with parallel actions or coordination. Our analysis also reveals LLMs' potential on abstract tasks where they can focus on high-level parallel optimization. ParaCook provides a scalable evaluation framework with adjustable complexity, establishing a ...
