I did a proper scientific experiment using RAWs published by various sources. Lightroom opens 90D CR3's without problems, which is nice.
Now, it is well-known that any properly prepared IQ test specimen should be underexposed at least five stops, and my sources have apparently failed in this task. All the photos I've found are more or less well-exposed. So we must settle for the next best thing and just pick a photo with shadows as deep as possible. After some deliberation I chose
this one.
Consider first Specimen A, an unprocessed upper left corner crop of the aforementioned picture.

As is known, the next step is to open the file in Lightroom and push the shadows slider all the way to 100, resulting in Specimen B:

This is clearly not enough, so next we shall produce Specimen C by adjusting exposure by +2 EV:

We're still seeing too much detail, and banding or fixed pattern noise is inconveniently absent. Let's push another two EV for Specimen D:

At shadows +100, exposure +4 EV, what we have looks… interesting. Definitely should have underexposed more. Still, we might as well push exposure all the way to +5 EV while we're at it and end up with Specimen E:

Is this good? Or is this bad? I do not know, for I am not a Sony fanboy.