Blog/What is an open-weight model?
Definition

What is an open-weight model?

By Samuel Seidel · Published September 9, 2026

An open-weight model is a machine learning model whose trained parameters, the weights, are published for anyone to download and run on their own hardware. The label describes distribution, not how the model was built: the training data, training code and evaluation method can all stay private while the weights are free to download.

Weights are the output, not the recipe

A model's weights are the numbers learned during training, the parameters that turn an input into an output once the architecture is fixed. Downloading a model's weights gets you a working copy you can run yourself, fine-tune, or inspect. It doesn't get you the dataset that produced those weights, the training scripts, the compute budget, or the decisions made about data filtering and curation along the way. Those are separate artifacts, and most open-weight releases don't include them.

Why this differs from "open source"

In software, open source means you can see and modify the source that produces the program, as well as the program itself. Applied loosely to models, "open source" often just means open-weight, which is a much weaker claim: you can run and fine-tune an open-weight model, but you usually can't reproduce it from scratch, audit what it was trained on, or verify claims about data provenance. Some organizations release training code and dataset details alongside the weights, which is closer to open source in the software sense; among current releases that is the exception.

Licensing is a separate axis from openness

A model being open-weight says nothing about what you're allowed to do with it. Licenses attached to open-weight releases range from permissive (Apache 2.0, MIT) to custom terms with commercial-use thresholds, redistribution limits, or acceptable-use restrictions tied to the model provider's own policies. Two open-weight models can sit on very different points of that spectrum, so the license text for the specific release matters more than the "open" label. For when open-weight is the right choice over a closed API, see the comparison of open-weight vs closed-weight models.

What open-weight makes possible

With the parameters in hand, you can run the model on hardware of your choosing, fine-tune it, quantize it to fit smaller hardware, or inspect its behavior directly instead of only through an API's output. A closed-weight model allows none of that: the provider keeps the parameters on its own servers and exposes only an API. That is why self-hosting is only possible with open-weight models; see the glossary for related terms like quantization and inference.

Checking a release before you rely on it

Before relying on a release, check three things on the model's own page or repository: the license terms of that specific release, whether the publisher discloses its training data sources, and whether independent benchmark results exist beyond the publisher's own. Releases vary widely in how much of this they document. A model card that says nothing about data provenance is common; note it before you build something that depends on the answer.

Where GPUwerk fits

If you want to run an open-weight model without buying hardware, a GPUwerk Spark is a dedicated DGX Spark with 128 GB of unified memory to load and serve it, from $0.79/hour in EU-Central. You choose the model, the quantization and the inference engine; no model API sits between you and the weights. More detail is on the private LLM hosting page.

Related pages

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

Run your own weights.

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

Create account