Glossary
Parameters (AI models)
The numerical values inside a model that are adjusted during training and determine its behavior.
Parameters, often called weights, are the numbers inside a neural network that are tuned during training. Together they encode what the model has learned. A model's size is usually described by how many parameters it has.
More parameters generally allow a model to learn more, but they also make it more expensive to train and run. Size is only one factor: data quality, training methods, and the amount of computing used also matter, and a smaller, newer model can outperform a larger, older one.
When a developer releases a model's weights, others can run and modify the model themselves. That is what "open-weight" means.