Fable 5 is back — Fable vs Sonnet 5, same app, one shot

Why watch

Matt Maher compares Fable 5 and Sonnet 5 AI coding assistants by giving each the same task: build a native macOS SwiftUI radial launcher app in a single attempt with no edits. The video introduces the Care Benchmark, which tests whether models correctly incorporate user requests into their planning phase rather than just optimizing for generic code quality. Results include bug analysis, cost breakdowns, and evaluation of whether Anthropic's /goal feature justifies its higher token usage.