I have studied this particular issue in absurd depth—waaaaay beyond just in vivo testing. In retrospect, there are probably better ways that I could have spent my time, so I hope my DBG glip-glops* can benefit from it.
Malinckrodt generic IR mph is garbage; some batches are awful, and no batches are better than "kinda-maybe barely acceptable." Watson-Actavis was the gold standard (literally exactly as good as brand), but their IR mph operations died after Teva took the reins. KVK is somewhat inferior to the old W-A stuff, but it's consistently fine. Sandoz mph seems to have vanished into the abyss, but it was comparable to KVK, albeit less suitable for any parenteral ROA due to insoluble excipients.
Personally, I'd have no objection to paying an extra hundred bucks or so to get a month's worth of brand, W-A , Sandoz, or KVK instead of Malinckrodt, but I probably wouldn't pay very much to get brand instead of Sandoz or KVK (maybe I'd pay ~$20 extra?). I wouldn't pay any extra at all to get brand instead of W-A.
Batch-to-batch variability is a real thing, especially among generics (and especially especially in the case of Malinckrodt), so people who haven't noticed a difference with generics may very well have been fortunate enough to have gotten exclusively good batches. In a large enough sample, across enough batches over enough time, my above remarks will hold in a statistical sense, but they shouldn't be construed to be at odds with anybody's small-sample experience. I.e., I'm very confident in my answer, but I'm not claiming that people who haven't noticed differences are necessarily wrong.
Hope this helps!
*Brownie points if you get the reference