mlsanityData to 14 Oct 2026

The Interaction Tax: When Communication Erases Diversity in Multi-Agent Teams

arXiv 2608.23541 ↗First posted 24 Aug 2026cs.MA cs.AIcited by 5

On the map: LLM agents and tool use → Social simulation and multi-agent debate

Abstract (arXiv)

Does multi-agent LLM interaction help or hurt? Some work reports gains from debate (Du et al., 2024), critique loops (Chen et al., 2025), and mixture-of-agents synthesis (Wang et al., 2025), while other work finds that interaction adds cost without improving quality under equal budgets (Tran & Kiela, 2026; Xu et al., 2026; Jarrett et al., 2025), or that independent sampling already captures multi-agent gains (Li et al., 2024). We argue this contradiction partly reflects a missing distinction, because not all multi-agent communication is equal. Different model families find structurally different solutions, but when agents read each other's complete outputs, their proposals converge within one round, erasing the diversity that motivates using multiple models. We call this the interaction tax. We test 11 verifier-scored optimization tasks under matched budgets and find that full-solution interaction is a weak default. Independent proposal generation avoids this collapse. Full-solution interaction mainly makes agents stay close to the first solution they see instead of trying different approaches, and critique helps only if the violated rule is easy for the LLM to find and fix. These results suggest that multi-agent performance depends less on the number of agents than on the information they exchange, and interaction helps only when agents share the right information at the right time.

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Placed in Social simulation and multi-agent debate, one of 16 directions in LLM agents and tool use. Builds on 8 of its 19 references that are in the map. First posted on arXiv 24 Aug 2026.

Builds on

8 of 19 references.

References · 8

  1. 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,926
  2. 2023Improving Factuality and Reasoning in Language Models through Multiagent Debatecited by 1,784
  3. 2023Plan-and-Solve Prompting: Improving Zero-Shot Chain-of-Thought Reasoning by Large Language Modelscited by 560
  4. 2024Mixture-of-Agents Enhances Large Language Model Capabilitiescited by 438
  5. 2024MAgICoRe: Multi-Agent, Iterative, Coarse-to-Fine Refinement for Reasoningcited by 41
  6. 2026Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgetscited by 30
  7. 2026Rethinking the Value of Multi-Agent Workflow: A Strong Single Agent Baselinecited by 22
  8. 2026Multi-Agent Teams Hold Experts Backcited by 21

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 5.

5 papers.

  1. 2026UnitBoost: Managing Compound LLM Systems with a Merge Operator, Not a Modelcited by 0
  2. 2026CollabFlow: Recursive Self-Improvement of Agent Collaborationcited by 0
  3. 2026Pay for the Fault, Not the Flow: Label-Free In-Flow Multi-Agent Workflow Optimizationcited by 0
  4. 2026AReaL-RSI: Adaptive Scaling and Coordination of Recursive Self-Improvementcited by 0
  5. 2026Scaling Discovery Requires Forgettingcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026The Collaboration Tax: How Much LLM Multi-Agent Systems Pay to Coordinate
  2. 2026When Agents Disagree: The Selection Bottleneck in Multi-Agent LLM Pipelines
  3. 2026Diversity Collapse in Multi-Agent LLM Systems: Structural Coupling and Collective Failure in Open-Ended Idea Generation
  4. 2026Beyond Symmetric Agents: Cognitive Diversity and Multi-Agent Debate in Small Language Models
  5. 2026More Capable, Less Cooperative? When LLMs Fail At Zero-Cost Collaboration
  6. 2026Towards a Science of Collective AI: LLM-based Multi-Agent Systems Need a Transition from Blind Trial-and-Error to Rigorous Science
  7. 2026Diversity Without Fidelity: A Solver-Sampler Mismatch in Multi-Agent LLM Negotiation Simulation
  8. 2026Attention Tax, Handoff Tax: A Stylised Model of When Multi-Agent LLM Systems Help