I'm not sure, maybe it's a paid feature, I'm on the $10 a month plan, which is Suno's cheapest next to free. I used their free one for about a year, but eventually upgraded cause it was worth it to me since I use it a lot - more credits than I usually use, but I sometimes I hate having to wait for the next day to generate something on the free plan, so I upgraded. Also, not that it's something I'm worried about but with the free plan you don't own your compilations, Suno does, and it's for only personal use. With the paid plan you own your creations and can distribute and/or whatever you want to use it for because you own it..
By the way, I said it gave me 24 channels because that's what it looked like it outputted, but it was actually 12 stems, what I saw was apparently the left and right channels of each stem, thus the 24. Oh yeah.. it cost 50 credits to extract the stems.
As for myself, if I were still trying to restore whatever vintage recordings, I would much prefer not using internet based processing (like Suno) to accomplish it. I would sooner use something like your UVR software based solely on my own computer. I wasn't really suggesting Suno would be better for your application, I just didn't really realize it could do that and thought it interesting enough to point out.
... And the process isn't perfect. Some of the audio (the really high notes) I found on the guitar track. The tracks that I didn't include in the final mix didn't have anything on them but noise.
And most of the piano ended up on the vocal track, with some on the piano track. Weird. I think it might be a byproduct of the acoustic recording. The model was probably trained on pristine, CD audio.
Anyway, it seemed to work fairly well.
I mentioned I extracted the lion roars from my track, one was about midpoint and the other roar at the end, what I found peculiar is that they came out on two different individual stems, it seems to me it should of been on the same one. And with the horns, they seem to be on two different tracks to.. it seems to be separating the - what - frequencies? But you don't necessarily get all the horns or guitar or what ever on a single channel... Not sure what I'm talking about but I can associate with what your saying above.