ModelChorus lets you compare AI models head-to-head in blind challenges. Submit a prompt, judge the responses, and help build a community-driven leaderboard.
Use a suggestion button or write your own prompt. When you submit, two anonymous models receive it and generate responses in real time.
Read both responses side-by-side. Judge quality, accuracy, style, and reasoning — without knowing which model wrote which.
Pick the response you prefer, or call it a tie. Your vote feeds into our Elo rating system, helping rank models across topics and languages.
After voting, the model names are revealed. Continue chatting in the same thread or start a fresh comparison with a new prompt.
Voting does more than rank models. Every challenge you judge earns points that unlock tiers, badges, and a place on the community leaderboard — your account keeps track of all of it.
Every challenge you judge earns you 10 points, up to 300 a day, so your score climbs with each comparison.
Your points fill a progress bar that carries you through six ranks, from Listener to Conductor.
Vote day after day to build a streak, with badges rewarding runs of up to 60 days.
Earn badges for vote milestones up to 2,500 votes, streak runs, and judging text, image, and audio.
Your votes earn you a place among the community on the voter leaderboard.
During occasional promotions, verified accounts can put points toward a limited number of gift cards and other rewards.
Every vote updates model rankings using the Elo rating system — the same method used in chess. Win streaks push models up; losses pull them down. The leaderboard reflects real community preferences.
We keep the evaluation process transparent while respecting your privacy. Review our policies for details on data handling.
