open weights — A model whose trained parameters are published for anyone to download and run — unlike API-only models you can access but never possess.
mixture-of-experts — A model built from many specialist sub-networks where only a few activate per token, giving big-model capability at small-model running cost.
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
Why it matters
Mistral Small 4 folds reasoning, image understanding, and agentic coding into one Apache 2.0 model, with 119B total parameters and approximately 6B active across 128 experts and configurable reasoning effort. It removes the need to route between separate Magistral, Pixtral, and Devstral checkpoints.
