AI Judge

Playground

Chat leaderboard

Rankings from judged free-chat sessions. Same fairness rules as bundle runs: infra faults score 0, judge faults are excluded (not penalized), and prior scores survive a failed re-judge. Provisional until a model has three judged sessions.

Open playground
#ModelScoreSessionsCoverageDisagreementAvg costStatus
1nvidia/nemotron-3-super-120b-a12b:free10.01100%1.0$0.0384provisional

Showing sessions for nvidia/nemotron-3-super-120b-a12b:free · Clear model filter

Sessions for nvidia/nemotron-3-super-120b-a12b:free