Okay, when I check usage from lxtask, I am also at 380MB, so at least we're getting pretty consistent readings when using the same tools on the same distro, but different rigs.
I have always used top as a benchmark, for consistency across distros and desktop environments. lxtask obviously calculates RAM usage differently than top or htop. But it hardly seems fair to use lxtask as the benchmark, as it's typically only available in LXDE installations, or someone specifically installed that package. Gnome, XFCE, KDE all have their own implementation "task" managers, which I'm guessing also calculate RAM usage differently. For that reason, IMO, lxtask isn't a reliable benchmark unless you're comparing LXDE to LXDE. Just my opinion.
When considering a new desktop environment awhile ago, I did several "benchmark" readings on various desktop environments, on different distros, for an overall comparison. Nothing definitively scientific, but general readings of RAM usage in default states from cold boots. Here's a table of those findings:
The only way to do this with any semblance of consistency was to use CLI tools, available on every distro. Again, certainly not definitively scientific, but it was absolutely close enough for me to assess what I was considering. By this metric, Manjaro LXDE would have placed very well on this table...no surprise, being a lightweight DE based on Arch, which generally scored better than every other distro I looked at. FWIW...