Top news
Investing in the public benefit OfficialAnthropic NewsOct 18Popular Musk Pledges SpaceX Shares to ‘Every Kid in America’ WebBloomberg TechnologyOct 19Popular Websites are closing their doors to AI agents WebThe Verge AIOct 17 This is the new ChatGPT Voice, powered by GPT Live OfficialOpenAI on YouTubeOct 18 How AI giants will reshape global philanthropy WebAxios AIOct 19 Gemini 4 Argon Really Surprised Me WebTwo Minute PapersOct 19 Google Gemini 4 agents just beat Minecraft with no human intervention WebEngadget AIOct 19 Who could it be? 🧐 OfficialGoogle on YouTubeOct 18 Anthropic gives outside AI watchdog “employee-like” access WebArs Technica AIOct 18 What are self-replicating ‘AI worms’, and why is Anthropic worried about them? WebThe Guardian AIOct 18
GlossaryAgentsIntermediate

Multi-agent system

Also written: MAS, Multi-agent orchestration, Multi-agent architecture, Multi-agent framework, Multi-agent AI system, Multi-agent teams, Multi-agent pipelines, Agent pipelines

Multiple AI agents collaborating to achieve a common goal, each using tools and interacting with each other and their environment.

In plain terms

Imagine a team of AI assistants, each with a specific skill, working together on a complex project. A multi-agent system is like that team. Instead of one AI trying to do everything, several specialized AI agents are coordinated to complete a task.

For example, one agent might be good at finding information online, another at summarizing it, and a third at writing a report. They pass information back and forth, using their tools and skills to reach the final goal, like creating a research paper or analyzing data.

The expert version

A multi-agent system is an artificial intelligence architecture composed of multiple autonomous agents that interact with each other and their environment to achieve a common objective. These agents, often large language models (LLMs), can leverage tools, communicate, and coordinate their actions. This approach allows for complex problem-solving by decomposing tasks, parallel processing, and specialization among agents, potentially leading to more robust and efficient outcomes than a single monolithic agent.

Tradeoffs include increased complexity in coordination, communication overhead, and potential for emergent behaviors that are difficult to predict or control. However, it enables sophisticated workflows like research assistance, complex data analysis, and interactive simulations.

Why it matters now

Frontier AI labs are increasingly exploring multi-agent systems to tackle complex tasks that are beyond the capabilities of a single AI model. Recent developments show a focus on improving the efficiency, reliability, and collaborative abilities of these AI teams, especially under constraints like latency and budget. This shift reflects a move towards more sophisticated AI architectures that can orchestrate multiple specialized agents.

Where it shows up 25 items, newest first
How each company uses it
Anthropic Claude
Anthropic is developing patterns for building multi-agent teams that can operate efficiently under latency and budget pressures, and is researching the effectiveness of these teams in tasks like social science research.
Google Gemini
Google Gemini is providing tools and tutorials for building reliable multi-agent systems, moving beyond simple prompts to more deterministic agent routing and orchestration.
Microsoft Copilot
Microsoft Copilot is integrating multi-agent concepts into its code review process, enabling automated resolution and analysis by AI agents.
OpenAI ChatGPT
OpenAI is involved in research and discussions surrounding AI safety and capabilities, which implicitly includes the development and implications of multi-agent systems.
Related terms

AgentLLMTool UseOrchestrationAI WorkflowAutonomous Agent

Often confused with
Agent
An agent is a single AI program designed to perform tasks, whereas a mult
LLM
A large language model is the underlying AI technology that can power indiv
Distributed Computing
This refers to systems where computation is spread across
Origin

The concept of multi-agent systems originates from research in artificial intelligence and distributed artificial intelligence, aiming to model and create systems composed of multiple interacting intelligent entities. Early work dates back to the 1980s, exploring coordination and negotiation among agents. Source ↗

Try this about 5 minutes

Ask an AI assistant like ChatGPT or Claude to act as multiple characters in a role-playing scenario. For instance, ask it to simulate a debate between a historian and a scientist on a specific topic. Observe how different 'agents' (personas) interact and present distinct viewpoints to achieve the goal of a simulated debate.

Try it: Playbook entries that use thisPlaybook →

The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.

Questions people ask

What is the main benefit of a multi-agent system?
The main benefit is the ability to tackle complex problems by breaking them down into smaller tasks that specialized agents can handle, leading to more efficient and robust solutions.
How do agents in a multi-agent system communicate?
Agents can communicate through various means, including passing messages, sharing data, or using a common communication language, depending on the system's design.

First seen on Subvolts Sep 5, 2024. Updated Oct 15, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.