One measure for code quality should be the effort to make a minor change to code, or to fix a simple, common bug.

Can a developer who is familiar with the code, effortless make a change in an area where change was expected at the beginning?

Assuming a typical, minor bug that turns out to originate from the usual suspect places in the code (database query, logical error, incomplete implementation of a requirement, off-by-one/calculation error). Is it usually easy to narrow things down, and to spot the error?

That means the lead architect should regularly try implementing minor features or fix minor bugs and/or pair up with colleagues who do, and draw conclusions about the code accordingly.