mlsanityData to 7 Oct 2026

Why Do Multi-Agent LLM Systems Fail?

arXiv 2503.13657 ↗First posted 17 Mar 2025cs.AIcited by 559

On the map: LLM agents and tool use → Software engineering coding agents

Abstract (arXiv)

Despite enthusiasm for Multi-Agent LLM Systems (MAS), their performance gains on popular benchmarks are often minimal. This gap highlights a critical need for a principled understanding of why MAS fail. Addressing this question requires systematic identification and analysis of failure patterns. We introduce MAST-Data, a comprehensive dataset of 1600+ annotated traces collected across 7 popular MAS frameworks. MAST-Data is the first multi-agent system dataset to outline the failure dynamics in MAS for guiding the development of better future systems. To enable systematic classification of failures for MAST-Data, we build the first Multi-Agent System Failure Taxonomy (MAST). We develop MAST through rigorous analysis of 150 traces, guided closely by expert human annotators and validated by high inter-annotator agreement (kappa = 0.88). This process identifies 14 unique modes, clustered into 3 categories: (i) system design issues, (ii) inter-agent misalignment, and (iii) task verification. To enable scalable annotation, we develop an LLM-as-a-Judge pipeline with high agreement with human annotations. We leverage MAST and MAST-Data to analyze failure patterns across models (GPT4, Claude 3, Qwen2.5, CodeLlama) and tasks (coding, math, general agent), demonstrating improvement headrooms from better MAS design. Our analysis provides insights revealing that identified failures require more sophisticated solutions, highlighting a clear roadmap for future research. We publicly release our comprehensive dataset (MAST-Data), the MAST, and our LLM annotator to facilitate widespread research and development in MAS.

Key facts. Cited by 559 papers in the map (data through 7 Oct 2026). Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use. Builds on 63 of its 176 references that are in the map. First posted on arXiv 17 Mar 2025.

Builds on

63 of 176 references.

Closest ancestors · 6

Same direction, newest first

  1. 2025Interactive Debugging and Steering of Multi-Agent AI Systems3 Mar
  2. 2024HyperAgent: Generalist Software Engineering Agents to Solve Coding Tasks at Scale9 Sep
  3. 2024AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents26 Jul
  4. 2024OpenHands: An Open Platform for AI Software Developers as Generalist Agents23 Jul
  5. 2023AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation16 Aug
  6. 2023ChatDev: Communicative Agents for Software Development16 Jul

Other references · 57

  1. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  2. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,393
  3. 2020Measuring Massive Multitask Language Understandingcited by 7,768
  4. 2023Generative Agents: Interactive Simulacra of Human Behaviorcited by 3,500
  5. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,383
  6. 2023SWE-bench: Can Language Models Resolve Real-World GitHub Issues?cited by 3,193
  7. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 3,155
  8. 2023A Survey on Large Language Model based Autonomous Agentscited by 2,203
  9. 2023Code Llama: Open Foundation Models for Codecited by 2,105
  10. 2023MetaGPT: Meta Programming for A Multi-Agent Collaborative Frameworkcited by 1,761
  11. 2023Improving Factuality and Reasoning in Language Models through Multiagent Debatecited by 1,739
  12. 2023The Rise and Potential of Large Language Model Based Agents: A Surveycited by 1,422
38 more
  1. 2023CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Societycited by 1,348
  2. 2023Gorilla: Large Language Model Connected with Massive APIscited by 1,302
  3. 2023MemGPT: Towards LLMs as Operating Systemscited by 1,295
  4. 2024Qwen2.5-Coder Technical Reportcited by 1,151
  5. 2023GAIA: a benchmark for General AI Assistantscited by 1,084
  6. 2021The Surprising Effectiveness of PPO in Cooperative, Multi-Agent Gamescited by 903
  7. 2023DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelinescited by 803
  8. 2024Large Language Model based Multi-Agents: A Survey of Progress and Challengescited by 752
  9. 2023ChatEval: Towards Better LLM-based Evaluators through Multi-Agent Debatecited by 724
  10. 2023Trustworthy LLMs: a Survey and Guideline for Evaluating Large Language Models' Alignmentcited by 394
  11. 2023A Survey on Large Language Model (LLM) Security and Privacy: The Good, the Bad, and the Uglycited by 380
  12. 2023Check Your Facts and Try Again: Improving Large Language Models with External Knowledge and Automated Feedbackcited by 351
  13. 2024Agent Workflow Memorycited by 342
  14. 2023Multi-Agent Collaboration: Harnessing the Power of Intelligent LLM Agentscited by 329
  15. 2022Large Language Models are Better Reasoners with Self-Verificationcited by 304
  16. 2024MT-Bench-101: A Fine-Grained Benchmark for Evaluating Large Language Models in Multi-Turn Dialoguescited by 247
  17. 2023Building Cooperative Embodied Agents Modularly with Large Language Modelscited by 231
  18. 2024Magentic-One: A Generalist Multi-Agent System for Solving Complex Taskscited by 221
  19. 2018Learning Attentional Communication for Multi-Agent Cooperationcited by 170
  20. 2024GSM-Plus: A Comprehensive Benchmark for Evaluating the Robustness of LLMs as Mathematical Problem Solverscited by 155
  21. 2025Multi-Agent Risks from Advanced AIcited by 150
  22. 2024Does Prompt Formatting Have Any Impact on LLM Performance?cited by 132
  23. 2024Exploring Large Language Model based Intelligent Agents: Definitions, Methods, and Prospectscited by 130
  24. 2018Learning when to Communicate at Scale in Multiagent Cooperative and Competitive Taskscited by 114
  25. 2023RLTF: Reinforcement Learning from Unit Test Feedbackcited by 110
  26. 2025Scaling Test-Time Compute Without Verification or RL is Suboptimalcited by 82
  27. 2024Prover-Verifier Games improve legibility of LLM outputscited by 79
  28. 2024AutoKaggle: A Multi-Agent Framework for Autonomous Data Science Competitionscited by 70
  29. 2023LLM-Coordination: Evaluating and Analyzing Multi-agent Coordination Abilities in Large Language Modelscited by 65
  30. 2024StateFlow: Enhancing LLM Task-Solving through State-Driven Workflowscited by 65
  31. 2024TestGenEval: A Real World Unit Test Generation and Test Completion Benchmarkcited by 49
  32. 2023Towards Reasoning in Large Language Models via Multi-Agent Peer Review Collaborationcited by 48
  33. 2024Optima: Optimizing Effectiveness and Efficiency for LLM-Based Multi-Agent Systemcited by 40
  34. 2023BenchMARL: Benchmarking Multi-Agent Reinforcement Learningcited by 31
  35. 2024A Survey of Useful LLM Evaluationcited by 28
  36. 2024BattleAgentBench: A Benchmark for Evaluating Cooperation and Competition Capabilities of Language Models in Multi-Agent Systemscited by 26
  37. 2014Social Learning and Distributed Hypothesis Testingcited by 20
  38. 2024TeamCraft: A Benchmark for Multi-Modal Multi-Agent Systems in Minecraftcited by 20

50 most cited of 57 All 57 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 160; 2026: 399.

559 papers.

Sort:

By area:

  1. 2025Towards a Science of Scaling Agent Systemscited by 125
  2. 2025Where LLM Agents Fail and How They can Learn From Failurescited by 118
  3. 2025AgenTracer: Who Is Inducing Failure in the LLM Agentic Systems?cited by 105
  4. 2025Multi-Agent Collaboration via Evolving Orchestrationcited by 102
  5. 2025TRAIL: Trace Reasoning and Agentic Issue Localizationcited by 98
  6. 2025ResearchRubrics: A Benchmark of Prompts and Rubrics For Evaluating Deep Research Agentscited by 80
  7. 2025Debate or Vote: Which Yields Better Decisions in Multi-Agent Large Language Models?cited by 74
  8. 2025Latent Collaboration in Multi-Agent Systemscited by 67
  9. 2025Multi-Agent Evolve: LLM Self-Improve through Co-evolutioncited by 58
  10. 2025MARFT: Multi-Agent Reinforcement Fine-Tuningcited by 51
  11. 2025LLM-based Agents Suffer from Hallucinations: A Survey of Taxonomy, Methods, and Directionscited by 49
  12. 2025Barbarians at the Gate: How AI is Upending Systems Researchcited by 48
38 more
  1. 2025Creating General User Models from Computer Usecited by 44
  2. 2025El Agente: An Autonomous Agent for Quantum Chemistrycited by 43
  3. 2026Code as Agent Harnesscited by 40
  4. 2025Vibe Coding vs. Agentic Coding: Fundamentals and Practical Implications of Agentic AIcited by 39
  5. 2025Measuring Agents in Productioncited by 35
  6. 2025Fara-7B: An Efficient Agentic Model for Computer Usecited by 34
  7. 2025Who is Introducing the Failure? Automatically Attributing Failures of Multi-Agent Systems via Spectrum Analysiscited by 33
  8. 2026Dive into Claude Code: The Design Space of Today's and Future AI Agent Systemscited by 33
  9. 2025Enabling Agents to Communicate Entirely in Latent Spacecited by 31
  10. 2026Formal Policy Enforcement for Real-World Agentic Systemscited by 31
  11. 2025KramaBench: A Benchmark for AI Systems on Data-to-Insight Pipelines over Data Lakescited by 30
  12. 2025MedAgentBoard: Benchmarking Multi-Agent Collaboration with Conventional Methods for Diverse Medical Taskscited by 29
  13. 2026From Agent Traces to Trust: A Survey of Evidence Tracing and Execution Provenance in LLM Agentscited by 29
  14. 2025Thought Communication in Multiagent Collaborationcited by 28
  15. 2026Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgetscited by 28
  16. 2025Agentic AI Security: Threats, Defenses, Evaluation, and Open Challengescited by 26
  17. 2025When Agents go Astray: Course-Correcting SWE Agents with PRMscited by 25
  18. 2025Stronger-MAS: Multi-Agent Reinforcement Learning for Collaborative LLMscited by 25
  19. 2026The Long-Horizon Task Mirage? Diagnosing Where and Why Agentic Systems Breakcited by 25
  20. 2025TAMAS: Benchmarking Adversarial Risks in Multi-Agent LLM Systemscited by 24
  21. 2025Process-Centric Analysis of Agentic Software Systemscited by 24
  22. 2025Creativity in LLM-based Multi-Agent Systems: A Surveycited by 23
  23. 2025Optimas: Optimizing Compound AI Systems with Globally Aligned Local Rewardscited by 23
  24. 2025DoVer: Intervention-Driven Auto Debugging for LLM Multi-Agent Systemscited by 23
  25. 2026ContextBench: A Benchmark for Context Retrieval in Coding Agentscited by 23
  26. 2025Large Language Model-based Data Science Agent: A Surveycited by 22
  27. 2025Emergent Coordination in Multi-Agent Language Modelscited by 22
  28. 2026Why Reasoning Fails to Plan: A Planning-Centric Analysis of Long-Horizon Decision Making in LLM Agentscited by 22
  29. 2025SABER: Small Actions, Big Errors -- Safeguarding Mutating Steps in LLM Agentscited by 21
  30. 2026Rethinking Failure Attribution in Multi-Agent Systems: A Multi-Perspective Benchmark and Evaluationcited by 21
  31. 2025Weak-for-Strong: Training Weak Meta-Agent to Harness Strong Executorscited by 20
  32. 2025Collaborating Action by Action: A Multi-agent LLM Framework for Embodied Reasoningcited by 20
  33. 2025AI Scientists Fail Without Strong Implementation Capabilitycited by 20
  34. 2025When Identity Skews Debate: Anonymization for Bias-Reduced Multi-Agent Reasoningcited by 20
  35. 2026CooperBench: Why Coding Agents Cannot be Your Teammates Yetcited by 20
  36. 2025RAFFLES: Reasoning-based Attribution of Faults for LLM Systemscited by 19
  37. 2025Aegis: Automated Error Generation and Attribution for Multi-Agent Systemscited by 19
  38. 2025Adaptation of Agentic AI: A Survey of Post-Training, Memory, and Skillscited by 19

50 most cited of 559 All 559 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026MAS-FIRE: Fault Injection and Reliability Evaluation for LLM-Based Multi-Agent Systems
  2. 2025Which Agent Causes Task Failures and When? On Automated Failure Attribution of LLM Multi-Agent Systems
  3. 2026SoK: When Safe Agents Fail Together: The Security of Multi Agent LLM Systems
  4. 2026When Agents Fail to Act: A Diagnostic Framework for Tool Invocation Reliability in Multi-Agent LLM Systems
  5. 2026Understanding Issues, Causes and Solutions in Open-Source LLM-based Multi-Agent Systems
  6. 2026Seeing the Whole Elephant: A Benchmark for Failure Attribution in LLM-based Multi-Agent Systems
  7. 2025How Do LLMs Fail In Agentic Scenarios? A Qualitative Analysis of Success and Failure Scenarios of Various LLMs in Agentic Simulations
  8. 2025MASLab: A Unified and Comprehensive Codebase for LLM-based Multi-Agent Systems