Parameters
Also called model weights
The billions of internal numbers a model adjusts during training; together they hold what it learned.
A model is, in the end, a very large set of numbers called parameters, or weights. Training nudges them, billions of times, until the model's predictions get good. Everything it "knows", from grammar to the boiling point of water, is spread across those numbers rather than stored as facts you could look up.
Parameter counts are how makers describe size: a few billion for models that run on a phone, hundreds of billions or more for the biggest. More parameters generally means more capability and more cost to run, but size alone doesn't decide quality. What the model was trained on, and how it was tuned afterwards, matter as much.
For example
OpenAI's 2 open-weight gpt-oss models have about 117 billion and 21 billion parameters; the smaller one runs on a computer with 16 GB of memory.
Put it to use
Where you'll meet parameters in ChatGPT, Gemini and Claude, how to use it there, and the mix-up most people make.
Daily Bytes Plus is in the Android app.