parameter
A parameter is an internal value of an AI model that is set through training, and the parameter count is commonly used to describe a model's size.
In a neural network, parameters mainly refer to the weights and biases of the connections between neurons. The size of a large language model is usually stated as its parameter count, so 27B means 27 billion parameters. More parameters require more memory and computation to run. In programming, the word also means a value passed to a function.
This entry is based on AIPOST articles and widely known facts. If something is wrong, please send us a correction request.
Articles covering this entry
Requests follow a strict schema, a fixed format, so the parameters are always valid and can be executed directly.
The model's weights and parameters did not change.
Gemma 4 is an open-weights model family, meaning Google publishes the trained parameters so anyone can download them.
Growing models to 10 trillion or 100 trillion parameters, the internal values a model adjusts during training, will not produce transparent reasoning on its own…
Overnight, the agent produced five or six working versions, gathered parameters for feedback and laid out seven possible next development steps.
Model weights are the stored parameters the system needs before it can generate an image.
Open-weight means a model’s parameters are available to users; the result describes this test, not every model or deployment.
An open-weight model makes its model parameters available for others to adapt.
In Qwen 3.8 27B Q4, Qwen identifies the model family, 3.8 identifies the release, and 27B means 27 billion parameters, or stored weight values.