ChatDevvsMetaGPT
Generated from the two spec rows. Green marks the better value where a spec has a clear direction. Everything else is just different.
ChatDev
OpenBMB · Agent framework
OSS
Panel
5.40 spec wins
- Reliability
- 4.8
- Usefulness
- 4.8
- Cost
- 7.0
- Longevity
- 4.8
“A virtual software company staffed entirely by language models, which is also the business plan of several real ones.”
MetaGPT
DeepWisdom · Agent framework
OSS
Panel
5.63 spec wins
- Reliability
- 4.7
- Usefulness
- 5.0
- Cost
- 7.0
- Longevity
- 5.8
“It simulates a whole software company, including the part where somebody writes a competitive analysis nobody reads.”
Spec by spec
| Spec | ChatDev | MetaGPT |
|---|---|---|
| Architecture | ||
| Category | Agent framework | Agent framework |
| Runs | local | local |
| Platforms | macos, linux, windows | macos, linux, windows |
| Context window | not documented | not documented |
| Protocols | ||
| MCP client | No | No |
| MCP server | No | No |
| Capabilities | ||
| Runs terminal commands | Yes | Yes |
| Multi-file edits | Yes | Yes |
| Git operations | No | No |
| Browser control | No | No |
| Sandboxed execution | No | No |
| Multi-agent orchestration | Yes | Yes |
| Headless / CI mode | No | Yes |
| Models | ||
| Backbone | any | GPT, Azure OpenAI, Ollama, Groq |
| Bring your own model | Yes | Yes |
| Local models | No | Yes |
| Cost | ||
| Pricing model | byok | byok |
| Starts at | $0/mo | $0/mo |
| Free tier | Yes | Yes |
| Bring your own key | Yes | Yes |
| Openness | ||
| Open source | Yes | Yes |
| License | Apache-2.0 | MIT |
| GitHub stars | 34,431 | 70,718 |
Which one would each critic pick
| Critic | ChatDev | MetaGPT | Pick |
|---|---|---|---|
| El Juez | — | — | not enough reviews |
| El Amigo | 5.8 | 5.8 | no preference |
| El Crítico | 5.8 | 4.8 | ChatDev — Version 2.0 turned a research project into a zero-code console and pushed the classic line to a legacy branch, so the version every paper describes is now the old one. |
| El Profesor | 4.8 | 5.3 | MetaGPT — Standard operating procedures encode a waterfall in which each role hands a written artefact to the next, and no stage is documented as validating the one before it. |
| La Inversora | 5.5 | 6.0 | MetaGPT — Seventy thousand stars against no product, no hosted service and no published pricing, which is one of the largest gaps between attention and revenue on this board. |
| La Jefa | 4.8 | 4.8 | no preference |
| El Hacker | 5.8 | 7.3 | MetaGPT — MIT, one YAML file holds the entire configuration, and it takes any OpenAI-compatible endpoint including my own, so the whole simulated company runs on my hardware. |
Picks are derived from each critic's own scores. Humans vote on matchups on the duels page.