Information
# AWorld: The Agent Runtime for Self-Improvement
*"Self-awareness: the hardest problem isn't solving within limits, it's discovering the own limitations"*
[](https://x.com/InclusionAI666)
[](https://raw.githubusercontent.com/inclusionAI/AWorld/main/readme_assets/aworld_wechat_qr.jpg)
[](https://discord.gg/b4Asj2ynMw)
[](https://opensource.org/licenses/MIT)
[](https://deepwiki.com/inclusionAI/AWorld)
[中文版](./README_zh.md)
## Table of Contents
- [News](#news) — Latest updates and announcements.
- [Introduction](#introduction) — Overview and purpose of the project.
- [Installation](#installation) — Step-by-step setup instructions.
- [Quick Start](#quick-start) — Get started with usage examples.
- [Architecture](#architecture) — Explore the multi-agent system design.
- [Demo](#demo) — See the project in action with demonstrations.
- [Contributing](#contributing) — How to get involved and contribute.
- [License](#license) — Project licensing details.
## News
- [2025/07/07] AWorld, as a runtime, is now ready for agentic training. See [Self-Improvement section](#self-improvement-with-diverse-runtimes) for details. We have updated our score to 77.08 on the GAIA test. Learn how to construct a GAIA runtime in the [Demo section](#demo-of-gaia-agent-runtime).
- [2025/06/19] We have updated our score to 72.43 on the GAIA test. Additionally, we have introduced a new local running mode. See \`./README-local.md\` for detailed instructions.
- [2025/05/22] For quick GAIA evaluation, MCP tools, AWorld, and models are now available in a single Docker image. See ./README-docker.md for instructions and [youtube video](https://www.youtube.com/watch?v=kkYWeVvJKrg) for demo.
- [2025/05/13] AWorld has updated its state management for browser use and enhanced the video processing MCP server, achieving a score of 77.58 on GAIA validation (Pass@1 = 61.8) and maintaining its position as the top-ranked open-source framework. Learn more: [GAIA leaderboard](https://huggingface.co/spaces/gaia-benchmark/leaderboard)
- [2025/04/23] AWorld ranks 3rd on GAIA benchmark (69.7 avg) with impressive Pass@1 = 58.8, 1st among open-source frameworks. Reproduce with python examples/gaia/run.py
## Introduction
AWorld (Agent World) is a multi-agent playground that enables agents to collaborate and self-improve. The framework supports a wide range of applications, including but not limited to product prototype verification, foundation model training and Multi-Agent System (MAS) design meta-learning.
### Runtime Key Features
| 1. Agent Construction | 2. Topology Orchestration | 3. Environments |
|----------------------|--------------------------|-----------------|
| • Support for various model services • Integration with MCP tools
• Custom tool support | • Protocol encapsulation between models and tools
• Protocol encapsulation among agents | • Runtime state management
• State tracing support
• Distributed, high-concurrency environments for training | ### Self-Improvement with Diverse Runtimes By constructing diverse runtime environments (with tools, agents, or models in them), AWorld aims to find the limitations of a model and push intelligence forward. Here we will record some of our work to prove the effectiveness of our proposal. | Category | Runtime | Performance | Key Information | |-----|----------------|-------------|--------------| | Tool Use | Function call runtime to be released | Competitive on BFCL benchmark
 | []()
[](https://huggingface.co/Bingguang/FunReason)
[](https://arxiv.org/pdf/2505.20192)
[]()
[](https://github.com/BingguangHao/FunReason)| | Deep Search | Search runtime to be released | SOTA on HotpotQA benchmark
 | [](https://github.com/inclusionAI/AgenticLearning)
[](https://huggingface.co/collections/endertzw/rag-r1-68481d7694b3fca8b809aa29)
[](https://arxiv.org/abs/2507.02962)
[](https://github.com/inclusionAI/AgenticLearning)| ### Demo of GAIA Agent-Runtime  Here we first introduce the **GAIA runtime**, which can be constructed on your local computer. It can be used for: - **Product prototype verification** - **Self-improvement training** (See [training pipeline](#backward) for details) Follow the instructions in [\`./examples/gaia/README.md\`](./examples/gaia/README.md) to initialize the GAIA agent runtime and run the demo shown above. > **Want to build your own multi-agent system? Check out the detailed tutorials below to get started! ** ## Installation Python>=3.11: \`\`\`bash git clone https://github.com/inclusionAI/AWorld cd AWorld python setup.py install \`\`\` ## Quick Start > Here's a quick start guide to: (1) create your first agent; (2) equip it with a MCP tool; (3) assign a teammate; and (4) answer a user query through teamwork. \`\`\`python from aworld.config.conf import AgentConfig from aworld.agents.llm_agent import Agent from aworld.runner import Runners from aworld.core.agent.swarm import Swarm if __name__ == '__main__': agent_config = AgentConfig( llm_provider="openai", llm_model_name="gpt-4o", # Set via environment variable or direct configuration # llm_api_key="YOUR_API_KEY", # llm_base_url="https://api.openai.com/v1" ) # Register the MCP tool here, or create a separate configuration file. mcp_config = \{ "mcpServers": \{ "amap-amap-sse": \{ "type": "sse", "url": "https://mcp.amap.com/sse?key=YOUR_API_KEY", "timeout": 5, "sse_read_timeout": 300 \} \} \} # Create your first agent equipped with an MCP tool search = Agent( conf=agent_config, name="search_agent", system_prompt="You are a helpful agent.", mcp_servers=["amap-amap-sse"], # MCP server name for agent to use mcp_config=mcp_config ) # Add a new teammate to the agent summary = Agent( conf=agent_config, name="summary_agent", system_prompt="You are a helpful summary agent." ) # Collaborate as a team; the default is a static workflow swarm = Swarm(search, summary) # Run agent team res = Runners.sync_run(input="Hotels within 1 kilometer of West Lake in Hangzhou", swarm=swarm) print(res) \`\`\` ## Architecture AWorld is designed to achieve two primary objectives: (1) provide an efficient forward process, and (2) facilitate diverse backward processes, including but not limited to foundation model training and system design meta-learning. ### Forward > An illustration of the runtime, showing the message workflow when Agent1 receives a query from a user.  ### Backward > During training, an action-state rollout demonstration using **AWorld's distributed environments**.  > An illustration of training code that seamlessly integrates the RL learning framework (Swift, in this example) with AWorld as the environment is shown below. This integration enables scalable and efficient agent training through distributed environment execution. (To run high-concurrency rollouts, you need to deploy an online distributed environment. Please contact [chenyi.zcy@antgroup.com](mailto:chenyi.zcy@antgroup.com) if assistance is needed.) #### Usage To apply and use this integration: 1. Clone AWorld's \`agent_training_server\` branch: \`\`\`bash git clone -b agent_training_server --single-branch https://github.com/inclusionAI/AWorld.git AWorld \`\`\` 2. Clone ms-swift's v3.5.2 branch (shallow clone): \`\`\`bash git clone -b v3.5.2 --depth=1 https://github.com/modelscope/ms-swift.git ms-swift \`\`\` 3. Copy patch files from AWorld to ms-swift: \`\`\`bash cp -r AWorld/patches ms-swift/ \`\`\` 4. Enter the patches directory and apply the patch: \`\`\`bash cd ms-swift/patches git apply 0001-feat-add-agent-training-support-with-aworld-server.patch \`\`\` ## Demo > Running Pre-defined Agents (e.g., see [demo code](examples/browsers/run.py)). Below are demonstration videos showcasing AWorld's capabilities across various agent configurations and environments.
| Mode | Type | Demo |
|---|---|---|
| Single Agent | Browser use |
Watch Browser Demo on YouTube |
| Phone use |
Watch Mobile Demo on YouTube |
|
| Multi Agent | Cooperative Teams |
Watch Travel Demo on YouTube |
| Competitive Teams |
Watch Debate Arena on YouTube |
|
| Mixed of both Teams | Coming Soon |
Reply