First impressions: Fable 5.1 less 'Claudish', Databricks dogfoods Kimi K3 daily

MaziyarPanahi · x · 2026-09-03

YuchenjUW's day-one take on Fable 5.1: more human-sounding, strong at frontend but web search lags GPT-5.6 Sol. His daily driver is now Kimi K3, which Databricks heavily dogfoods and finds surprisingly good.

Related event: Anthropic Launches Claude Fable 5.1 With 75% Cheaper Cache but Contested Real-World Costs(168 posts)→

Original post →

More from Models

Models channel →