The article presents an updated version of the SWISS-MODEL server, which has been continuously developed over the past 25 years to become a leading tool in automated protein homology modeling. The new features include the ability to model homo- and heteromeric protein complexes, a new modeling engine (ProMod3), and an improved local model quality estimation method (QMEANDisCo). The server uses evolutionary information to infer the stoichiometry and overall structure of complexes from the amino acid sequences of interacting proteins. The performance of the updated SWISS-MODEL is evaluated through the CAMEO project, where it consistently ranks among the top modeling servers for various aspects of protein structure prediction. The article also provides a case study on the modeling of the Ferredoxin-Ferredoxin-NADP(+) Reductase complex, demonstrating the effectiveness of the new features in generating accurate and biologically meaningful models.The article presents an updated version of the SWISS-MODEL server, which has been continuously developed over the past 25 years to become a leading tool in automated protein homology modeling. The new features include the ability to model homo- and heteromeric protein complexes, a new modeling engine (ProMod3), and an improved local model quality estimation method (QMEANDisCo). The server uses evolutionary information to infer the stoichiometry and overall structure of complexes from the amino acid sequences of interacting proteins. The performance of the updated SWISS-MODEL is evaluated through the CAMEO project, where it consistently ranks among the top modeling servers for various aspects of protein structure prediction. The article also provides a case study on the modeling of the Ferredoxin-Ferredoxin-NADP(+) Reductase complex, demonstrating the effectiveness of the new features in generating accurate and biologically meaningful models.