Wait, I finally found a paper that properly presents the results in a way that lets you decide for yourself! See figure 2 here: https://pmc.ncbi.nlm.nih.gov/articles/PMC9344377/#f2, you want to focus on the second plot with the dotted and dashed lines.

Basically, both placebo and drugs can have huge responses, but basically, for severe depression, yes, it looks like the drug is more likely to have a large positive effect than the placebo is likely to have a large positive effect.

There are still a lot of limitations, main one being that study criteria generating the data for this would exclude most cases of really high severity actually seen in practice. Also "blinding" is mostly BS in these kinds of studies, and antidepressants have massive, noticeable effects, so the proper comparison is really an active placebo, and we really can't be sure the whole apparent antidepressant effect is basically just "woah, I can really feel this doing shit", basically, at some level, and this is still bad because the placebo vs. drug difference is still super tiny and below what we would consider typical minimal clinically important differences.

So still hard to say. But, yeah, there is a good chance the drugs are more effective in severe cases, and, technically, you should definitely treat in a severe case.