From my experience, performance gains seen in Debug builds in Unity/C#/Mono nearly always translate in gains in Release mode. I know that this is not always true, but in this context that's my experience.

Setting up release benchmarks is much more complex and we develop the game in Debug mode, so it is very natural to get the first results there, and if promising, validate them in Release.

Also, since our team works in Debug mode, even gains that only speed things up in Debug mode are valuable for us, but I haven't encountered a case where I would see 20%+ perf gain in Debug mode that would not translate to Release mode.

I agreed with you initially, but is it really as big of a deal in C#? I'm used to compiled languages where "debug build" means "no compiler optimisations", aka every operation done with a variable is a memory load + store, trivial functions aren't inlined so even trivial accessors called in a loop carry function call overhead, etc etc. But this is C#, so the JIT presumably optimises just as much in a debug build as in a release build?

So in C++ terms, it's really just benchmarking "-O2" instead of "-O2 -DNDEBUG". This seems fine.