Rethinking Multi-Agent Collaboration: When More Is Less
Yishuo Yuan and colleagues at Shanghai Jiao Tong University and Nanjing University map where multi-agent collaboration beats a single-agent harness and propose SAIGE, which models collaboration as a semantic dependency graph that grows on demand.
Ask this paper
The boundary is task structure, not scale. Multi-agent collaboration helps on long-horizon tasks with sparse dependencies; single-agent harnesses stay better on tightly coupled sequential workflows.
Collaboration is formalised as a dependency DAG. Trajectories are decomposed into subtasks via bridge edges, and orchestration becomes bridge-edge prediction, which gives an explicit account of the context cost each split saves or adds.
SAIGE spawns nodes on demand. Agent instances are created as needed and edges encode semantic dependencies established by content-based retrieval, rather than fixing a topology in advance.
More agents stop helping. Scaling the agent pool or deepening recursion does not consistently improve outcomes, which is the measured form of the paper's title claim.
Failures are traced to edge misprediction. A case study shows a mispredicted bridge edge producing a local deadlock, which identifies the component that limits the method.
Abstract
The rapid advancement of large language models and single-agent harnesses has reshaped the landscape of autonomous systems, raising a critical question of when multi-agent collaboration offers genuine value. As individual agent capabilities continue to scale, multi-agent collaboration faces diminishing returns while incurring growing context overhead. Through systematic analysis, we delineate the capability boundaries of multi-agent collaboration relative to single-agent alternatives, showing that it confers systematic benefits specifically in long-horizon tasks with sparse dependencies, while single-agent harnesses remain superior in tightly coupled, sequential workflows. Building on these insights, we propose SAIGE, a lightweight multi-agent collaboration mechanism based on Semantic-Aware Incremental Graph Evolution. SAIGE models collaboration as a dynamically evolving graph, where nodes are agent instances spawned on demand and edges encode semantic dependencies established through content-based information retrieval. Experiments on long-horizon, complex task benchmarks show that SAIGE achieves a favorable trade-off between context efficiency and task performance, and that scaling the agent pool or deepening the recursion level does not consistently improve outcomes. Our findings suggest that multi-agent superiority is bounded by task structure rather than universal, and that more agents do not necessarily make a system more intelligent.