AI Leaderboard
Repos ranked by AI analysis. Multi-model scores are averaged — expand cards to compare opinions.
...
Rank by
How it works
Composite Score
Average of coolness, innovation, and practicality across all models that scored the repo.
Multi-Model Consensus
Consensus = models agree (σ < 0.8). Mixed = some disagreement. Split = strong disagreement.
Independent Scoring
Each model scores independently — they never see other models' scores. Expand cards to compare their reasoning.