# Mixture of Experts **Entity class:** Model architecture **As of:** November 17, 2026 ## Definition **Mixture of Experts** is a model architecture routing work to selected expert components, relevant to expert parallelism when experts sit on different accelerators. This node belongs to the **distributed inference** route of [[wiki/Distributed Compute Ecosystem|Distributed Compute Ecosystem]], which records the specific kind of work or network capacity involved. ## Relationships - **comparison:** [[wiki/Megatron-LM|Megatron-LM]]. - **mechanism route:** [[wiki/Distributed Compute Ecosystem|Distributed Compute Ecosystem]]. - **source article:** [[articles/Embedded Super-Spreader Events and Muse|Embedded Super-Spreader Events and Muse]]. ## Sources - [Mixture of Experts — project record or attributed account](https://docs.nvidia.com/megatron-core/developer-guide/latest/user-guide/parallelism-guide.html).