Jason Droege

We’re introducing SEAL Showdown, the AI leaderboard that actually captures real preferences, powered by a platform used by real people. Public benchmarks today rely on contrived tasks or narrow user groups. That leaves us guessing which models are actually preferred by people. SEAL Showdown changes that. Model performance can now be segmented by demographics and domains. Rankings aren’t just a single global average, they can be broken down by region, profession, education level, age and more, giving a nuanced view of how models work for different people. SEAL Showdown sets a new bar, because AI should be judged by how well it works for everyone.