Soient donn\'es deux graphes \(\Gamma_1\), \(\Gamma_2\) \`a \(n\) sommets. Sont-ils isomorphes? S'ils le sont, l'ensemble des isomorphismes de \(\Gamma_1\) \`a \(\Gamma_2\) peut \^etre identifi\'e avec une classe \(H \pi\) du groupe sym\'etrique sur \(n\) \'el\'ements. Comment trouver \(\pi\) et des g\'en\'erateurs de \(H\)? Le d\'efi de donner un algorithme toujours efficace en r\'eponse \`a ces questions est rest\'e longtemps ouvert. Babai a r\'ecemment montr\'e comment r\'esoudre ces questions -- et d'autres qui y sont li\'ees -- en temps quasi-polynomial, c'est-\`a-dire en temps \(\exp(O(\log n)^{O(1)})\). Sa strat\'egie est bas\'ee en partie sur l'algorithme de Luks (1980/82), qui a r\'esolu le cas de graphes de degr\'e born\'e. English translation: Graph isomorphisms in quasipolynomial time [after Babai and Luks, Weisfeiler--Leman,...]. Let \(\Gamma_1\), \(\Gamma_2\) be two graphs with \(n\) vertices. Are they isomorphic? If any isomorphisms from \(\Gamma_1\) to \(\Gamma_2\) exist, they form a coset \(H \pi\) in the symmetric group on \(n\) elements. How can we find a representative \(\pi\) and a set of generators for \(H\)? Finding an algorithm that answers such questions efficiently (in all cases) is a challenge that has long remained open. Babai has recently shown how to solve these problems and related ones in quasipolynomial time, i.e., time \(\exp(O(\log n)^{O(1)})\). His strategy is based in part on an algorithm due to Luks (1980/82), who solved the case of graphs of bounded degree.