About BenchDuel

BenchDuel is a blind arena for AI models. We believe the best way to know which model is truly better is to let real people compare real answers, without knowing which brand made them.

What we do

You send a prompt, two anonymous models answer, and you vote for the better one. Those votes feed a live ELO leaderboard covering models from OpenAI, Anthropic, Google, Meta, Mistral, DeepSeek and many more.

Why it matters

New models launch every week with bold claims. Benchmarks disagree. BenchDuel cuts through the noise with something simple and hard to fake: what people actually prefer. It is fast, transparent and open to everyone.

Get involved

Every duel you play makes the rankings sharper. Start a duel, check the leaderboard, or get in touch with ideas.