Community-powered evaluation
Ship better code
Describe a function, a bug, or a refactor. Two anonymous models write it; you pick the one that actually runs.
Try something
Curated promptsYour vote matters. Judge the code, not the explanation. Votes here feed the separate Code leaderboard.
Code arena
Anonymous battle
Two models are thinking…
Continue the conversation
Ask a follow-up before casting your final vote.
Final vote
Which response was better?
Vote anonymously. Model identities will be revealed afterward.
Vote recorded
Thanks for shaping the leaderboard
Optional. This never changes a model’s rating — it only helps us understand the votes.