| title | SWE-Team |
|---|---|
| emoji | 🤝 |
| colorFrom | red |
| colorTo | yellow |
| sdk | gradio |
| sdk_version | 5.50.0 |
| app_file | app.py |
| hf_oauth | true |
| pinned | false |
| short_description | Track GitHub team management statistics for SWE assistants |
SWE-Team ranks software engineering assistants by their real-world GitHub membership event activity.
No benchmarks. No sandboxes. Just real membership events tracked from public repositories.
Most AI coding assistant benchmarks use synthetic tasks and simulated environments. This leaderboard measures real-world activity: how many membership events is the assistant generating? How active is it across different projects? Is the assistant's usage growing?
If an assistant is consistently generating membership events across different projects, that tells you something no benchmark can.
Key metrics from the last 180 days:
Leaderboard Table
- Assistant: Display name of the assistant
- Website: Link to the assistant's homepage or documentation
- Total Membership Events: Number of team membership changes (e.g., adding or removing members) performed by the assistant
Monthly Trends
- Membership event volume over time (bar charts)
- Activity patterns across months
We focus on 180 days to highlight current capabilities and active assistants.
Data Collection We mine GitHub activity from GHArchive, tracking:
- Membership events by the assistant (
MemberEventdata)
Regular Updates Leaderboard refreshes daily
Community Submissions
Anyone can submit an assistant. We store metadata in SWE-Arena/bot_metadata and results in SWE-Arena/leaderboard_data. All submissions are validated via GitHub API.
Planned improvements:
- Repository-based analysis (which repos are assistants managing)
- Extended metrics (membership event types, member roles, access levels)
- Organization and team breakdown
- Member management patterns (invitations, removals, role changes)
Open an issue for bugs, feature requests, or data concerns.