Large Language Model (LLM)-based agentic systems have shown strong capabilities across various tasks. However, existing multi-agent frameworks often rely on static or task-level workflows, which either over-process simple queries or underperform on complex ones, while also neglecting the efficiency-performance trade-offs across heterogeneous LLMs. To address these limitations, we propose Difficulty-Aware Agentic Orchestration (DAAO), which can dynamically generate query-specific multi-agent workflows guided by predicted query difficulty. DAAO comprises three interdependent modules: a variational autoencoder (VAE) for difficulty estimation, a modular operator allocator, and a cost- and performance-aware LLM router. A self-adjusting policy updates difficulty estimates based on workflow success, enabling simpler workflows for easy queries and more complex strategies for harder ones. Experiments on six benchmarks demonstrate that DAAO surpasses prior multi-agent systems in both accuracy and inference efficiency, validating its effectiveness for adaptive, difficulty-aware reasoning.

翻译：基于大语言模型（LLM）的智能体系统已在多种任务中展现出强大能力。然而，现有的多智能体框架通常依赖于静态或任务级别的工作流，这要么对简单查询过度处理，要么在复杂查询上表现不佳，同时还忽视了异构大语言模型之间的效率-性能权衡。为解决这些局限性，我们提出了难度感知的智能体编排（DAAO），它能够根据预测的查询难度动态生成查询特定的多智能体工作流。DAAO包含三个相互依赖的模块：一个用于难度估计的变分自编码器（VAE）、一个模块化算子分配器，以及一个成本与性能感知的大语言模型路由器。一个自调整策略会根据工作流的成功情况更新难度估计，从而为简单查询启用更简单的工作流，并为更困难的查询启用更复杂的策略。在六个基准测试上的实验表明，DAAO在准确性和推理效率方面均超越了先前的多智能体系统，验证了其在自适应、难度感知推理方面的有效性。