A-MapReduce: Executing Wide Search via Agentic MapReduce

TL;DR

A-MapReduce introduces a parallel, MapReduce-inspired multi-agent framework for efficient wide search tasks in large language model systems, significantly improving performance and reducing computation time.

Contribution

It recasts wide search as a horizontally structured retrieval problem and implements a novel parallel processing framework with experiential memory for large-scale search tasks.

Findings

Achieves state-of-the-art performance on WideSearch benchmarks.

Reduces running time by 45.8% compared to baseline methods.

Delivers 5.11% - 17.50% average Item F1 improvements.

Abstract

Contemporary large language model (LLM)-based multi-agent systems exhibit systematic advantages in deep research tasks, which emphasize iterative, vertically structured information seeking. However, when confronted with wide search tasks characterized by large-scale, breadth-oriented retrieval, existing agentic frameworks, primarily designed around sequential, vertically structured reasoning, remain stuck in expansive search objectives and inefficient long-horizon execution. To bridge this gap, we propose A-MapReduce, a MapReduce paradigm-inspired multi-agent execution framework that recasts wide search as a horizontally structured retrieval problem. Concretely, A-MapReduce implements parallel processing of massive retrieval targets through task-adaptive decomposition and structured result aggregation. Meanwhile, it leverages experiential memory to drive the continual evolution of query-conditioned task allocation and recomposition, enabling progressive improvement in large-scale wide-search regimes. Extensive experiments on five agentic benchmarks demonstrate that A-MapReduce is (i) high-performing, achieving state-of-the-art performance on WideSearch and DeepWideSearch, and delivering 5.11% - 17.50% average Item F1 improvements compared with strong baselines with OpenAI o3 or Gemini 2.5 Pro backbones; (ii) cost-effective and efficient, delivering superior cost-performance trade-offs and reducing running time by 45.8\% compared to representative multi-agent baselines. The code is available at https://github.com/mingju-c/AMapReduce.