Before judging
- Entry ID
- GitHub handle when permission is granted
- Public challenge title or "Withheld until judging"
- Registration status
- Validation status
- Evaluation status
AgentRE Challenge / Leaderboard
Season 1 Public Leaderboard
Public status only during submissions. Private repository URLs, binaries, source code, ground truth, detailed techniques, interim scores, and full model transcripts are not displayed.
Entries
Registration status before judging, final difficulty scores after review.
During submissions, the public leaderboard may show entry ID, GitHub handle when permitted, challenge title when permitted, registration status, validation status, and evaluation status. Average correctness, difficulty score, and model outcomes are published only after judging and sanitization.
Results are locked during judging.
Validation and evaluation status may change while AgentRE reviews entries, but interim scores, model results, and full transcripts remain private until final publication.
No public entries yet.
Registered entries will appear here after public-safe registration PRs are merged. Private repositories, binaries, source code, ground truth, and repository URLs are never added to this page during submissions.
| Entry ID | Creator | Challenge Title | Registration | Validation | Evaluation |
|---|
Optional public repository links may be added after Season 1 only when explicitly approved by the participant.
Publication Model
What appears here changes by competition phase.
After judging
- Rank
- Creator
- Challenge title
- Models tested
- Average official model correctness
- Median official model correctness
- Official model panel wins
- Models below the passing threshold
- Complete failures
- Entry difficulty score
- Award or badge
- Sanitized challenge summary
- Optional public repository link after participant approval