What it is
In a neural network, model weights are values that are adjusted during the training process. They represent the strength of the connections between neurons, essentially encoding the knowledge and patterns learned from the training data. When new input data is fed into the model, these weights are applied to perform calculations, ultimately leading to the model's prediction or output.
Model weights are central to an AI model's functionality and performance. For "open-weight models," their public availability enables wider innovation and research, but also raises security and ethical considerations. The value of a proprietary AI model is often tied to its unique, highly optimized weights, which are the result of extensive training and compute. Their size and complexity continue to grow with model scale.
Why it matters
Model weights are the "brain" of an AI model; understanding them helps grasp how AI learns and why proprietary weights are valuable assets.
Reviewed under editorial standardsUpdated September 26, 2026Not investment advice