Dev: GPT models lead at computer use, Anthropic far behind despite Fable's coding lead

CtrlAltDwayne · x · 2026-09-03

A developer argues that even if Fable may be the leading coding model, it isn't the best choice for SWE work. He says computer use has become essential to his workflow and that nothing comes close to GPT models' computer-use capabilities, with Anthropic falling well behind. The post highlights the gap between benchmark-leading coding models and real-world agent usability.

Original post →

More from Models

Models channel →