Glossary
Active parameters
In a mixture-of-experts model, the subset of parameters actually used for each token. Inkling 2 Small has 594B total but about 17B active.
Sparse models route each token to a few of many expert sub-networks, so only a fraction of the weights do work on any given token. Total parameters set memory requirements and, loosely, knowledge capacity; active parameters set per-token compute, latency and serving cost. Comparing models by total parameters alone is misleading when one is sparse and the other dense.
Related terms
Mixture of experts (MoE)Open-weight modelParametersTrainingWeights
Get the brief in your inbox
One email every morning: the stories that mattered in AI, each with a one-line "why it matters" for your role. Free, no ads, unsubscribe any time.
Published daily by AI Briefing. We never share your address.