6Systems of agentsContested
Orchestrator and workers
A lead model splits a task as it goes, hands the parts to workers and combines their results.
Also called Orchestrator-worker pattern, Manager pattern, Supervisor pattern, Agents as tools
Anthropic, OpenAI, Cognition, LangChain
The problem
The subtasks are not known until the work begins.
This workflow is well-suited for complex tasks where you can’t predict the subtasks needed (in coding, for example, the number of files that need to be changed and the nature of the change in each file likely depend on the task).
Therefore
In the orchestrator-workers workflow, a central LLM dynamically breaks down tasks, delegates them to worker LLMs, and synthesizes their results.
The manager pattern empowers a central LLM—the “manager”—to orchestrate a network of specialized agents seamlessly through tool calls.
The supervisor pattern is a multi-agent architecture where a central supervisor agent coordinates specialized worker agents. This approach excels when tasks require different types of expertise.
In use
Claude Research
Our Research system uses a multi-agent architecture with an orchestrator-worker pattern, where a lead agent coordinates the process while delegating to specialized subagents that operate in parallel.
Claude Research, on Anthropic’s own evaluation
We found that a multi-agent system with Claude Opus 4 as the lead agent and Claude Sonnet 4 subagents outperformed single-agent Claude Opus 4 by 90.2% on our internal research eval.
Where it breaks
It uses many more tokens, and workers that cannot see each other’s work decide differently.
In our data, agents typically use about 4× more tokens than chat interactions, and multi-agent systems use about 15× more tokens than chats.
This is a tempting architecture, especially if you work in a domain of tasks with several parallel components to it. However, it is very fragile.
Sources
- Building effective AI agentsAnthropic, Dec 19, 2024
- A practical guide to building agentsOpenAI
- How we built our multi-agent research systemJeremy Hadfield, Barry Zhang, Kenneth Lien, Florian Scholz, Jeremy Fox, and Daniel Ford, Anthropic, Jun 13, 2025
- Don’t Build Multi-AgentsWalden Yan, Cognition, Jun 12, 2025
- Build a personal assistant with subagentsLangChain
The supervisor pattern is a multi-agent architecture where a central supervisor agent coordinates specialized worker agents.
- Agent orchestration - OpenAI Agents SDKOpenAI
Use agents as tools when a specialist should help with a bounded subtask but should not take over the user-facing conversation.
Added Nov 14, 2026