Users Claim Opus 5 Regresses Versus Fable Model

Peter Yang shared Kun Chenguid's critique noting Anthropic's Claude Opus 5 leads public benchmarks yet lags in actual tasks. Jeremy Howard switched back to Fable after testing, calling the newer model a regression. Yam Peleg amplified the report. The discussion centers on the gap between benchmark scores and real-world performance for Opus 5, with experienced users preferring the prior option based on direct experience.