LEADERBOARD

Real-time verification scores across frontier AI models. The BRIDGE Score measures how reliably a model communicates uncertainty, catches errors, and avoids hallucination across legal, medical, and code domains.

BRIDGE SCORE FORMULA

Agreement 40% + Catch Rate 25% + Inverse Miss 20% + Inverse Hallucination 15%
Last updated: --
# Model Agreement Catch Rate Inv. Miss Overall Trend

VERIFY YOUR MODEL

Submit a verification request and see how your model performs against the BRIDGE protocol in real-time.

Run a Verification →