lagent vs Qwen-Agent
Python frameworks for building LLM agents compared
lagent and Qwen-Agent are Python frameworks for building LLM-powered applications that use tools. lagent emphasizes composable agent components, memory, and multi-agent workflows, while Qwen-Agent focuses on Qwen-based assistants with integrations such as MCP, document retrieval, and code execution.

lagent: Build LLM Agents with a PyTorch-Inspired Framework
InternLM’s Lagent is a lightweight Python framework for building LLM-based agents, using a PyTorch-inspired design for composing agents and message passing. It supports synchronous and asynchronous workflows, tools, memory, and multi-agent applications.

Qwen-Agent: Build Tool-Using LLM Applications
Qwen-Agent is a Python framework for building Qwen-based assistants that can call tools, plan tasks, and work with documents. Use it to prototype or deploy agent applications when you need integrations such as MCP, code execution, or retrieval-augmented generation.
| lagent | Qwen-Agent | |
|---|---|---|
| Language | Python | Python |
| License | Apache-2.0 | Apache-2.0 |
| Stars | 2.3k | 17.1k |
| Forks | 248 | 1.7k |
| Last analyzed | Oct 15, 2026 | Oct 3, 2026 |
Key differences
- lagent offers component patterns for single agents and multi-agent workflows; Qwen-Agent centers on Qwen-based assistants and tool-using chat applications.
- lagent highlights structured agent messages, session memory, customizable aggregators, and output parsers; Qwen-Agent highlights function calling, parallel tool calls, and registering custom tools.
- Qwen-Agent includes examples for MCP, document question answering with RAG, and a Docker-based code interpreter; lagent's examples include browsing, code execution, and iterative writing workflows.
- lagent describes synchronous and asynchronous variants for many components; Qwen-Agent describes optional Gradio GUI support for trying an agent through a web interface.
- Both projects use Python and Apache-2.0 licensing. The supplied repository figures list 2.3k stars and 245 forks for lagent, and 17.1k stars and 1.7k forks for Qwen-Agent.
Choose lagent if you…
- need composable agent components with session memory and customizable message handling.
- want synchronous or asynchronous interfaces and patterns for multi-agent workflows.
- plan to build workflows such as research, code execution, or iterative writing with configurable model backends.
Choose Qwen-Agent if you…
- want to build assistants around Qwen models using DashScope or a compatible endpoint.
- need examples or integrations for MCP, document retrieval, parallel function calls, or a Docker-based code interpreter.
- want an optional Gradio interface for trying an agent through a web interface.
Compare the numbers side by side, or add more projects
This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.