Early Opus 5 feedback says Claude’s writing is now “4o-level slop,” despite stronger intelligence

jdjohnson · x · 2026-07-25

An X user says their first impression of Opus 5 is that Claude’s writing has become “slop” at roughly the same level as GPT-4o, even if the model remains smart underneath. They describe Fable as worth paying for because of the intelligence behind the output, but say they are unsure Opus itself is worth it.

Original post →

More from Models

Models channel →