# On the Reliability Limits of LLM-Based Multi-Agent Planning

arXiv: https://arxiv.org/abs/2603.26993 · HTML: https://mlsanity.com/paper/2603.26993/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use.
- Builds on 12 of its 30 references that are in the map.
- First posted on arXiv 27 Mar 2026.

## Placed in

- LLM agents and tool use / [General LLM agent frameworks and benchmarks](https://mlsanity.com/agents/general-llm-agent-frameworks-and-benchmarks/)

## Builds on (most cited first, 12 of 12)

- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [WebArena: A Realistic Web Environment for Building Autonomous Agents](https://mlsanity.com/paper/2307.13854/) (2023)
- [AgentBench: Evaluating LLMs as Agents](https://mlsanity.com/paper/2308.03688/) (2023)
- [OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments](https://mlsanity.com/paper/2404.07972/) (2024)
- [GAIA: a benchmark for General AI Assistants](https://mlsanity.com/paper/2311.12983/) (2023)
- [Why Do Multi-Agent LLM Systems Fail?](https://mlsanity.com/paper/2503.13657/) (2025)
- [Agentless: Demystifying LLM-based Software Engineering Agents](https://mlsanity.com/paper/2407.01489/) (2024)
- [Smart "Predict, then Optimize"](https://mlsanity.com/paper/1710.08005/) (2017)
- [From Predictive to Prescriptive Analytics](https://mlsanity.com/paper/1402.5481/) (2014)
- [Towards a Science of Scaling Agent Systems](https://mlsanity.com/paper/2512.08296/) (2025)
- [Talk Isn't Always Cheap: Understanding Failure Modes in Multi-Agent Debate](https://mlsanity.com/paper/2509.05396/) (2025)
- [Rethinking the Value of Multi-Agent Workflow: A Strong Single Agent Baseline](https://mlsanity.com/paper/2601.12307/) (2026)

## Built on by (most cited first, 8 of 8)

- [Coordination as an Architectural Layer for LLM-Based Multi-Agent Systems](https://mlsanity.com/paper/2605.03310/) (2026)
- [Biological Motifs for Agentic Control](https://mlsanity.com/paper/2607.04240/) (2026)
- [Semantic Register Compression in Multi-Agent LLM Cascades](https://mlsanity.com/paper/2607.14119/) (2026)
- [When Do Multi-Agent Systems Help? An Information Bottleneck Perspective](https://mlsanity.com/paper/2607.16133/) (2026)
- [Epi-Logic: A Conceptual Framework for Epistemic Runtime Control, Schema Validity Checking, and Controlled Accommodation in Autonomous AI Agents](https://mlsanity.com/paper/2609.24755/) (2026)
- [Re-derivability Decides What a Staged Agent Pipeline Recovers After an Upstream Fault](https://mlsanity.com/paper/2609.32802/) (2026)
- [Reliability Engineering for AI Systems: Challenges, Methods, and Directions](https://mlsanity.com/paper/2609.35316/) (2026)
- [Attention Tax, Handoff Tax: A Stylised Model of When Multi-Agent LLM Systems Help](https://mlsanity.com/paper/2610.06069/) (2026)

## Abstract (arXiv)

This technical note studies the reliability limits of LLM-based multi-agent planning as a delegated decision problem. We model the LLM-based multi-agent architecture as a finite acyclic decision network in which multiple stages process shared model-context information, communicate through language interfaces with limited capacity, and may invoke human review. We show that, without new exogenous signals, any delegated network is decision-theoretically dominated by a centralized Bayes decision maker with access to the same information. In the common-evidence regime, this implies that optimizing over multi-agent directed acyclic graphs under a finite communication budget can be recast as choosing a budget-constrained stochastic experiment on the shared signal. We also characterize the loss induced by communication and information compression. Under proper scoring rules, the gap between the centralized Bayes value and the value after communication admits an expected posterior divergence representation, which reduces to conditional mutual information under logarithmic loss and to expected squared posterior error under the Brier score. These results characterize the fundamental reliability limits of delegated LLM planning. Experiments with LLMs on a controlled problem set further demonstrate these characterizations.