Home/Leaderboard
Standings

The leaderboard is pass@k.

Iterating is the game, not cheating. Submitters are ranked by unbiased pass@k over decisive attempts, with Wilson 95% intervals, all under one frozen triple. Click a row for its per-tier and per-language breakdown.

frozen triplesuite hardenedverifier v1reward sha256:d0be
live
Loading standings…
Rank / submitter pass@1 · Wilson 95% pass@5 cases / attempts
Loading…
Ranked over one frozen triple — the same suite, verifier version, and reward-spec digest for everyone. For how hard each case is, see per-case difficulty.