Active parameters
The parameters a mixture-of-experts model actually uses for each token, usually a small share of its total.
Also: activated parameters
A dense model uses every parameter for every token. A mixture-of-experts model does not, so makers publish two numbers. Mixtral 8x7B has about 47 billion parameters in total but about 13 billion active per token. DeepSeek-V3 has 671 billion total and 37 billion active, and Kimi K2 about one trillion total and 32 billion active. Active parameters largely set the compute cost of each token, while total parameters set how much memory is needed to hold the model.