What Is a Closed-Source Model?

A closed-source model is an AI model whose internal weights and code are kept private by the company that made it, and used only through a paid interface or API. You can send it inputs and get outputs, but you cannot see, download, or run the model yourself.

How a closed-source model works

With a closed-source model, the provider hosts the model and exposes it through an API. You send a request and receive a response, while the weights, training data, and architecture stay with the provider. Widely used examples include the models behind ChatGPT, Claude, and Gemini. Access is usually metered and billed by usage, and you depend on the provider for availability and updates.

Closed-source vs open-source models

The difference is access and control. A closed-source model is easy to use and often high-performing, but you cannot host it, inspect it, or run it offline, and you rely on the vendor. An open-source model can be downloaded and run on your own infrastructure, which gives control and ownership at the cost of doing more of the work yourself. The choice usually comes down to control and data sensitivity against convenience.

At Custom AI Studio, the choice between closed and open models is made per system, based on control, cost, and how sensitive the data is.