Blog/What is a model weight?
Definition

What is a model weight?

By Samuel Seidel · Published September 9, 2026

A model weight is a single number stored inside a neural network. Each weight scales how much one piece of information affects the next step of a calculation as data moves through the network's layers. A large language model is, physically, a very long list of these numbers, arranged into matrices, plus the code that knows how to multiply them together in the right order.

Where the numbers come from

Weights start out random. Training adjusts them repeatedly: the model makes a prediction, the prediction is compared against the correct answer, and an algorithm called backpropagation nudges each weight slightly in the direction that would have made the prediction better. Repeated over enormous amounts of text and many passes, this turns a random list of numbers into one that can complete sentences, answer questions or write code. See inference vs training for how that process differs from actually running the model afterward.

What "parameter count" is counting

When a model is described as having, say, 8 billion parameters, that figure is a count of its weights. Not every weight matters equally to every output, but no subset is counted separately as "important"; the figure is the total number of learned values in the network. See what is a parameter count for how that number relates to capability and hardware requirements.

Precision: how many bits per weight

Each weight is stored using some number of bits, most commonly 16-bit floating point in the original training format. Reducing that to 8-bit or 4-bit representations, a process called quantization, shrinks the file size and the memory needed to run the model, generally at some cost to output quality. A 7-billion-parameter model at 16-bit precision needs roughly 14 GB of memory just to hold the weights; the same model at 4-bit precision needs closer to 4 GB. See what is model quantization for the mechanics and trade-offs.

Weights vs the rest of a model release

A model release usually bundles the weights with a tokenizer (which converts text to and from the numbers the model actually processes), a configuration file describing the network's architecture, and sometimes a license file. The weights are by far the largest part and the part that determines behavior; the rest is what you need to load and run them correctly. A model described as "open-weight" has these weight files published for anyone to download and run, which says nothing about whether its training data or method are open. See what is an open-weight model for that distinction.

Why weights matter for self-hosting

Because weights are files, an open-weight model can be downloaded once and run on your own or dedicated hardware, with no dependency on an external API staying available or unchanged. That's the basis of self-hosted inference: the weights sit on local storage, get loaded into GPU memory, and the machine you own or rent answers every request, with no third-party model API in the path.

A GPUwerk Spark is a dedicated DGX Spark for this pattern: download the weights once, then serve requests against them on the same machine, at $0.79/hour for one Spark or $1.79/hour for a linked pair. See the hardware page for memory specifications relevant to which weight sizes fit.

Related pages

First top-up: pay $20, get $30 in credit

Load the weights you chose on a machine nobody else rents.

One dedicated Spark, $0.79/hour, deployed in minutes in Europe, North America or East Asia.

Create account