KOL Slams Claude Opus 5 as Shallow and Inferior to GPT 5.6

andrewgwils · x · 2026-07-26

AI KOL @andrewgwils expressed shock at the poor performance of Claude Opus 5 so far. He reported issues including shallow understanding, weird mannered writing, and poor instruction following, bluntly stating that regardless of benchmarks, GPT 5.6 performs better in actual use.

Original post →

More from Models

Models channel →