Interesting, this is quite a stark difference in accuracy. I think this tool could prove very useful in the era of AI assistants.

Agree -- maintainable code was always important, but even more so in the age of AI. Unless we introduce proper guardrails via reliable code quality metrics, all these AI assistants will serve more as legacy code generators than genuine help.

A new benchmarking study shows CodeScene's Code Health metric is 6 times more accurate than SonarQube's software maintainability metric.