Closed Model is an AI model whose inner workings and trained parameters are kept private by the company that built it. You can typically use it only through a paid API or product, not download or modify it.
What it means in plain English
Closed (or “proprietary”) models are the opposite of open source ones. The company controls the model entirely — you access its capabilities through their service, but you cannot inspect its internals, run it on your own hardware, or change it. Many of the most capable AI models are closed, offered as polished, easy-to-use services in exchange for less transparency and control.
The choice between closed and open models is a key decision for anyone building with AI.
A simple example
When you use a leading AI assistant through its app or API, you are using a closed model: it works well and is easy to access, but you cannot see how it was built or run it independently of the provider.
Why it matters
Closed models offer convenience, polish, and often top performance, at the cost of transparency, control, and independence. The tension between closed and open models is one of the defining strategic and ethical debates in the AI industry.
Related terms
- Open Source Model — the transparent, self-hostable counterpart.
- API — how closed models are usually accessed.
- Foundation Model — many closed models are foundation models.
Frequently asked questions
What is a closed AI model?
A closed (or proprietary) model is one whose weights and often training details are not publicly released; you typically access it only through the provider’s API or product.
Closed vs open model — which is better?
It depends: closed models can offer top performance and managed convenience, while open models offer transparency, control, and self-hosting. The right choice depends on your needs around cost, privacy, and customisation.
