StreamMA：多智能体推理中的流式通信
Streaming Communication in Multi-Agent Reasoning
StreamMA 采用“流式通信”范式，每个推理步骤生成后立即流式传输给下游智能体，通过流水线相邻智能体降低端到端延迟。该方法还提升了效果，因为早期步骤更可靠，可避免错误后期步骤误导下游智能体。在数学、科学和代码八项推理基准上，使用 Claude Opus 4.6 和 GPT-5.4 两种大语言模型，及 Chain、Tree、Graph 三种拓扑，StreamMA 平均优于基线 +7.3 个百分点，在 HMMT 2026 上最高达 +22.4 个百分点。研究还发现“步骤级缩放定律”：增加每智能体步骤数可同时提升效果与效率。
让多 Agent 一边想一边传，不仅快了一倍还更准，这种流式思路要改写 pipeline 设计了，做多智能体的该认真读读。
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.
来源：HuggingFace Daily Papers（社区热门论文） · arxiv.org