Various real-life applications, for example, Internet of Things, wireless sensor networks, smart grids, transportation networks, communication networks, social networks, and computer grid systems, are always modeled as network structures. The binary-state network composed of binary-state (e.g., functioning or failed) components (arcs and/or nodes) is one of the most popular network structures. The two-terminal network reliability is a success probability that the network is still functioning and can be calculated by verifying the connectivity between two specific nodes, and is an effective and popular technique for evaluating the performance of all types of networks. To obtain complete information for a making better decisions, a multi-terminal network reliability extends the two specific nodes to a specific node subset in which all nodes are connected. In this study, a new algorithm called the all-multiterminal BAT is proposed by revising the binary-addition-tree algorithm (BAT) and the layered-search algorithm (LSA) to calculate all multi-terminal reliabilities. The efficiency and effectiveness of the proposed all-multiterminal BAT are analyzed from the perspective of time complexity and explained via numerical experiments to solve the all-multiterminal network reliability problems.