Thermal plasmas are utilized in diverse applications that require high power densities or throughputs, such as metal cutting, welding, spraying, metallurgy, and materials synthesis. Thermal plasma applications involve interactions between the highly energetic plasma and working gas streams, confining devices, or processing materials. Whereas thermal plasma implies a state of equilibrium (i.e. Local Thermodynamic Equilibrium, LTE), due to the above interactions, thermal plasma flows depict nonequilibrium phenomena of two types: kinetic and dissipative. Kinetic nonequilibrium manifests microscopically and is caused by localized imbalances between particles and fields interactions. Its occurrence is evidenced, for example, as deviations from thermal equilibrium between heavy-species and electrons or from mass-action laws. In contrast, dissipative nonequilibrium reveals macroscopically and is produced by external driving forces that incite distributed responses, such as the growth of instabilities, the occurrence of self-organization, or the establishment of turbulence. Although kinetic nonequilibrium has been increasingly incorporated in thermal plasma flow models (e.g. finite-rate chemistry, two-temperature models), it is the great advances in numerical computing that is enabling the exploration of dissipative nonequilibrium (e.g. pattern formation, small-scale turbulent features). Both types of nonequilibrium are reviewed, including their estimation and incidence, within the context of computational models and their relevance to thermal plasma sources and processes.