AutoGen
Microsoft's multi-agent conversation framework. Define multiple LLM-backed agents (with distinct system prompts, tools, and capabilities) and let them converse, plan, and execute tasks together. The biggest framework for orchestrating teams of specialised agents — code-writer + code-reviewer + executor patterns are the canonical use case. Backend-agnostic: works against any OpenAI-compatible endpoint, which includes most local runtimes (Ollama, vLLM, llama.cpp server). Python-first, with a v0.4 rewrite that split the codebase into autogen-core / autogen-agentchat / autogen-ext layers for cleaner extension.
Overview
Microsoft's multi-agent conversation framework. Define multiple LLM-backed agents (with distinct system prompts, tools, and capabilities) and let them converse, plan, and execute tasks together. The biggest framework for orchestrating teams of specialised agents — code-writer + code-reviewer + executor patterns are the canonical use case. Backend-agnostic: works against any OpenAI-compatible endpoint, which includes most local runtimes (Ollama, vLLM, llama.cpp server). Python-first, with a v0.4 rewrite that split the codebase into autogen-core / autogen-agentchat / autogen-ext layers for cleaner extension.
Pros
- Most mature multi-agent framework — battle-tested
- Works with any OpenAI-compatible local endpoint
- Studio UI for non-Python workflows
- Strong code-execution + tool-use patterns out of the box
Cons
- Multi-agent debugging is genuinely hard — chain explosions common
- v0.4 API differs sharply from v0.2 — community is split across both
- Token cost compounds fast (every agent's turn is a full inference call)
Compatibility
| Operating systems | linux macos windows |
| GPU backends | cuda rocm metal cpu |
| License | Open source · free |
Runtime health
Operator-grade signals on how actively AutoGen is being maintained, how fresh its measurements are, and what failure classes operators have flagged. Every label below is anchored to a real date or count — we never infer maintainer activity we can't show.
Release cadence
Derived from the most recent editorial signal on this row.
1 days since last refresh · source: lastUpdated
Benchmark freshness
How recent the editorial measurements on this runtime are.
No editorial benchmarks for this runtime yet.
Community reproduction
Submissions that match an editorial measurement on similar hardware.
No community reproductions on file yet.
Get AutoGen
Frequently asked
Is AutoGen free?
What operating systems does AutoGen support?
Which GPUs work with AutoGen?
Reviewed by RunLocalAI Editorial. See our editorial policy for how we evaluate tools.
Related — keep moving
Verify AutoGen runs on your specific hardware before committing money.