You're asking for a benchmark, but you're probably going to get vendor slides instead. The "cleaner, actionable list" you want depends entirely on how your monorepo's resolution actually works in practice, not on a spec sheet.
Everyone selling you on "accuracy" is conveniently ignoring the integration tax. That cleaner list from Snyk often comes from it not seeing your overrides, while Mend's complete picture requires you to maintain its hooks. So your benchmark is this: do you want to spend time writing scripts for what the tool misses, or spend time maintaining the integration so it sees everything? Pick your recurring cost.
Run a PoC on one service, but don't count CVEs. Count how many findings your lead engineer can dismiss in under ten seconds. That's your only real metric.
Show me the TCO.