lagent vs autogen
Python frameworks for LLM agents compared
lagent and autogen are Python frameworks for building LLM-powered agents and workflows that use models and tools. lagent emphasizes composable agent components, memory, and synchronous or asynchronous workflows; autogen offers layered APIs and prototyping tools for multi-agent applications, but is in maintenance mode.
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.
autogen: Build Multi-Agent AI Applications
AutoGen is a Python framework for building AI agents that work independently or alongside people. It supports multi-agent orchestration, model and tool integrations, and prototyping, but is now in maintenance mode.
| lagent | autogen | |
|---|---|---|
| Language | Python | Python |
| License | Apache-2.0 | CC-BY-4.0 |
| Stars | 2.3k | 61.3k |
| Forks | 248 | 9.3k |
| Last analyzed | Oct 15, 2026 | Oct 3, 2026 |
Key differences
- lagent provides components for structured messages, session memory, parsing, and tool execution; autogen includes Core, AgentChat, and Extensions APIs for agent messaging, orchestration, and integrations.
- lagent supports synchronous and asynchronous variants across many components, while autogen describes event-driven agents and local or distributed runtimes.
- autogen includes AutoGen Studio for no-code workflow prototyping and AutoGen Bench for evaluation; lagent's examples cover workflows such as code execution, browsing, and iterative writing.
- lagent uses the Apache-2.0 license; autogen uses CC-BY-4.0.
- lagent is not archived; autogen is in maintenance mode, with contributions limited to bug fixes, security patches, and documentation improvements.
- autogen's project ecosystem includes Python and .NET resources, while lagent is described as a Python framework.
Choose lagent if you…
- need composable Python components for agent messages, memory, parsing, and tool execution.
- want synchronous or asynchronous interfaces for single-agent or multi-agent workflows.
- are building workflows such as code execution, research, or iterative writing.
Choose autogen if you…
- need layered APIs for event-driven agents, higher-level chat patterns, and model integrations.
- want to prototype multi-agent workflows with AutoGen Studio or explore evaluation with AutoGen Bench.
- are maintaining an existing AutoGen project or prototyping, rather than starting a new project.
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.