Plagiarism of profound and significant results in contemporary mathematics research seems to be close to nonexistent. It’s antithetical to the culture of the discipline and the likelihood of success is remote when the spotlight is bright. Around the more dimly lit edges, however, it can and does happen from time to time as attested to in the comments.
I’ve encountered it myself. In 2012, after publishing some results in matrix analysis, I contacted an academic who was doing research on similar topics who I thought might be interested in some of my results. I summarized my results and we had a brief exchange and our correspondence ended not long afterwards. A few years later, in 2015, I attached some appendices to my full-text version of the 2012 paper at researchgate.net to help assist readers of the paper with illustrative examples and related MATLAB code. Upon posting it, I noticed that someone at the same Spanish university as the academic’s co-author had almost immediately downloaded the updated full-text. Then, later, in 2017, I came across one of their publications that took the same novel perspective, had nontrivial mathematical overlap, and similar MATLAB code generation results as my 2012 paper with appendices. Notably, there was no citation for my 2012 paper listed among their references. Instead, they (Lebtahi and Thome) had trotted out their results as if they had come up with it all on their own, apparently assuming I wouldn’t notice this violation of academic integrity. I sent an email requesting an explanation, and after more than a year and a half of attempting to work something out with them (and mostly being ignored), I’ve largely accepted that no meaningful action will be taken on their part, and attached yet another appendix to the article full-text detailing the transgression as I see it [if interested, you can look over Appendix E by accessing the full-text link accessible from here]. As to what their motivation was, I can’t say for sure. Was it tied to career advancement? A struggle for tenure? The idea that they could lift results from a relatively obscure journal without offering credit and get away with it?
At any rate, unlike priority disputes (which still happen from time to time), accusations of plagiarism that relate to the central core of mathematical research seem to be largely a relic of the past. Perhaps the most famous mathematical plagiarism accusations were hurled during the acrimonious debate between Newton and Leibniz on the development of calculus. Although a number of historical accounts suggest that this development was independent on the part of both parties, there are several indications that Leibniz was less than forthcoming about his knowledge of the research results obtained by others. One fascinating discussion about this appears in the excoriating book review that Brian Blank gave in the May 2009 AMS Notices (Volume 56, Number 5) for the book “The Calculus Wars”. As one example, on page 609, Blank points to evidence that suggests Leibniz’s Tentamen de Motuum Coelestium Causis was based on his uncredited reading of Newton's Principia. This book review is superbly written, and provides interesting details about this historic mathematical event. For those interested in the calculus controversy, I highly recommend checking out the review.