StreamMA通过流式传输推理步骤，降低延迟并提升准确性。
Streaming Communication in Multi-Agent Reasoning
- 推理步骤生成后立即传递，实现流水线处理。
- 在8个基准上平均提升7.3个百分点，最高达22.4个百分点。
- 适合需要高效高精度推理的多智能体系统应用。
多智能体推理系统采用‘生成-传输’范式，导致端到端延迟随流水线深度线性增长。我们提出StreamMA，一种将每个推理步骤在生成后立即传递给下游智能体的系统，通过流水线化相邻智能体来降低延迟。令人意外的是，这种流水线化还提升了有效性：因为多步推理质量非均匀，早期步骤比后期更可靠，使用这些可靠早期步骤而非完整推理链，可避免错误晚期步骤误导下游智能体。我们首次对流式、串行和单协议进行闭式联合分析，推导出有效性排序、加速上限和成本比。在涵盖数学、科学和代码的8个推理基准上，使用两个前沿大模型（Claude Opus 4.6 和 GPT-5.4）及三种拓扑（链式、树形、图状），StreamMA优于基线（平均+7.3个百分点，HMMT 2026最高+22.4个百分点；Claude Opus 4.6-high）。此外，我们发现‘步骤级缩放定律’：增加每智能体推理步骤数持续提升有效性和效率，这一新缩放维度与智能体数量缩放正交且可组合。
原文摘要 · Abstract (English)
Multi-agent reasoning systems adopt a "generate-then-transfer" paradigm that forces end-to-end latency to scale linearly with pipeline depth. We introduce StreamMA, a multi-agent reasoning system that streams each reasoning step to downstream agents as soon as it is generated, pipelining adjacent agents and thus reducing latency. Surprisingly, this pipelining also improves effectiveness: because multi-step reasoning quality is non-uniform and early steps are more reliable than later ones, working with these reliable early steps instead of the full chain prevents error-prone late steps from misleading downstream agents. We formalize both advantages with the first closed-form joint analysis of stream, serial, and single protocols, deriving the effectiveness ordering, speedup upper bound, and cost ratio. Across eight reasoning benchmarks spanning mathematics, science, and code, two frontier LLMs (Claude Opus 4.6 and GPT-5.4), and three topologies (Chain, Tree, Graph), StreamMA outperforms both baselines (avg. +7.3 pp, max +22.4 pp on HMMT 2026; Claude Opus 4.6-high). Beyond these contributions, we discover a "step-level scaling law": increasing per-agent steps consistently improves both effectiveness and efficiency, a new scaling dimension orthogonal to and composable with agent-count scaling.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。