#evaluations

  1. METR: AI is roughly doubling the speed of AI progress at Anthropic

    METR, the independent AI evaluator, spent 11 weeks inside Anthropic with unprecedented access to internal data.

  2. SUS-Bench: AI models beat people at a Mafia-style game

    A leaderboard from a 10-player social-deduction game (like Mafia or Among Us) where AI models play with and against real people, nobody…