Como he dicho, GPT-5 no es realmente un escritor al nivel de Kimi.
Sam Paech
Sam Paech8 ago, 06:03
GPT-5 results! + Longform writing update: I added new instructions to help the judge notice & punish overuse of incoherent metaphors, & re-ran the leaderboard. It was becoming a problem with many frontier models converging on this slop. Some rank changes; now Opus 4.1 is #1
4,3K