---
source: https://iinventedlight.com/library/notes/ao-handoff-orchestrator-agents.md
title: "Orchestrator agents: Magentic-One and the LLM-agent landscape"
license: CC-BY-4.0
license_url: https://creativecommons.org/licenses/by/4.0/
source_license: "CC BY 4.0; restated in our own words"
attribution: "Summary by iinventedlight (AI-assisted). Sources: arXiv, [2411.04468] Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks, CC BY 4.0; arXiv, [2308.11432] A Survey on Large Language Model based Autonomous Agents, CC BY 4.0."
author: "iinventedlight (synthesis of cited sources); text drafted with AI assistance"
date: 2026-11-04
status: published
checked_on: 2026-11-04
as_of: 2026-11-04
confidence: medium
untested: "No experiment was run; only the fetched arXiv abstract pages were read, so results are as reported by the authors."
category: agent-operations
topic_area: "agent-ops: multi-agent handoff"
tags: ['agents', 'multi-agent', 'orchestration', 'surveys']
sources:
  - id: S1
    title: "[2411.04468] Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks"
    publisher: arXiv
    url: https://arxiv.org/abs/2411.04468
    accessed: 2026-11-04
    license: "CC BY 4.0"
  - id: S2
    title: "[2308.11432] A Survey on Large Language Model based Autonomous Agents"
    publisher: arXiv
    url: https://arxiv.org/abs/2308.11432
    accessed: 2026-11-04
    license: "CC BY 4.0"
---
# Orchestrator agents: Magentic-One and the LLM-agent landscape

**Answer:** In Magentic-One, a lead agent called the Orchestrator plans, tracks progress, and re-plans to recover from errors, directing specialized agents to operate a web browser, navigate local files, or write and execute Python code (S1). A comprehensive survey delivers a systematic review of the field of LLM-based autonomous agents (S2).

## Magentic-One (S1)

- Agents must effectively plan, perform multi-step reasoning and actions, respond to novel observations, and recover from errors.
- Magentic-One achieves statistically competitive performance to the state-of-the-art on GAIA, AssistantBench, and WebArena.
- It achieves these results without modification to core agent capabilities or to how they collaborate.
- Its modular design allows agents to be added or removed from the team without additional prompt tuning or training.
- It ships with AutoGenBench, a standalone tool for agentic evaluation with built-in controls for repetition and isolation.

## The survey (S2)

- Previous research often trained agents with limited knowledge within isolated environments.
- Through the acquisition of vast amounts of web knowledge, LLMs have demonstrated remarkable potential in achieving human-level intelligence.
- It reviews applications in social science, natural science, and engineering, and the evaluation strategies commonly used.
- The authors maintain a repository of relevant references.

## Design notes

The modular design of Magentic-One allows agents to be added or removed, easing development (S1).

## Claims
- id: C1
  claim: The Orchestrator plans, tracks progress, and re-plans.
  source: S1
  evidence: "the Orchestrator, plans, tracks progress, and re-plans to recover from errors."
- id: C2
  claim: Magentic-One is open-source.
  source: S1
  evidence: "We provide an open-source implementation of Magentic-One"
- id: C3
  claim: The survey proposes a unified framework.
  source: S2
  evidence: "for which we propose a unified framework that encompasses a majority of the previous work."

## Sources
- [S1] [2411.04468] Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks. arXiv. https://arxiv.org/abs/2411.04468 . License: CC BY 4.0. Accessed 2026-11-04.
- [S2] [2308.11432] A Survey on Large Language Model based Autonomous Agents. arXiv. https://arxiv.org/abs/2308.11432 . License: CC BY 4.0. Accessed 2026-11-04.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Summary by iinventedlight (AI-assisted).