Multi-Frey= c n has no nontrivial positive integer solutions with (a, b) = 1 via a combination of techniques based upon the modularity of Galois representations attached to certain -ޑcurves, corresponding surjectivity results of Ellenberg for these representations, and extensions of multi-Frey curve arguments of Siksek.
IntroductionFollowing the proof of Fermat's last theorem [Wiles 1995], there has developed an extensive literature on connections between the arithmetic of modular abelian varieties and classical Diophantine problems, much of it devoted to solving generalized Fermat equations of the shapein coprime integers a, b, and c, and positive integers p, q, and r . That the number of such solutions (a, b, c) is finite, for a fixed triple ( p, q, r ), is a consequence of [Darmon and Granville 1995]. It has been conjectured that there are in fact at most finitely many such solutions, even when we allow the triples ( p, q, r ) to vary, provided we count solutions corresponding to 1 p + 2 3 = 3 2 only once. Being extremely optimistic, one might even believe that the known solutions constitute a complete list, namely (a, b, c, p, q, r ) corresponding to 1 p + 2 3 = 3 2 , for p ≥ 7, and to nine other identities (see [Darmon and Granville 1995;Beukers 1998]