Artificial Analysis Quantifies Claude Fable 5.2's Shift to Blunter Style

Artificial Analysis reports that Claude Fable 5.2 replies are 43% shorter with em-dash usage down 99%, and its AA-PersonalityBench shows the model became far more blunt, with its informativeness score dropping by 49 points versus 5.1.

2026-11-17 ~ 2026-11-17 · 2 related posts