Proprietary model: A proprietary model is an AI model whose weights are kept private by the company that built it. You use it through the company's app or API, and you cannot download, inspect, or self-host it.
OpenAI's GPT models, Anthropic's Claude, and Google's Gemini are proprietary. The vendor trains the model, runs it on its own servers, and sells access by subscription or by usage. You send a prompt over the internet, the vendor's servers generate the reply, and you get the text back.
The upside is that the strongest models at any given moment are usually proprietary, and the vendor handles hosting, scaling, safety tuning, and updates. The downside is dependence. The vendor can change pricing, retire a model version you built a workflow around, or alter behavior in an update. Your data also leaves your network, which puts the vendor's data-handling terms at the center of any compliance review.
In practice, most professionals use proprietary models through ChatGPT, Claude, Gemini, or Microsoft Copilot, and that is fine for everyday work. The decision gets more serious when you build a product or an internal tool on top of one. Then you want to know the vendor's retention policy, whether your inputs are used for training, what the service commitments are, and how hard it would be to swap in a different model later.
Example at work
A marketing agency standardizes on a paid team plan for a proprietary chat assistant because the admin controls, single sign-on, and no-training-on-your-data terms satisfy its client contracts. When the vendor retires the model version the agency's prompt library was tuned on, the team spends a week re-testing its templates.
Why it matters
Knowing whether a tool runs on a proprietary or an open-weights model tells you who controls your data, who sets the price, and what happens when the model changes. Ask those three questions before you make any AI tool part of a process your team depends on.
Related terms
- Open-weights modelAn open-weights model is an AI model whose trained parameters (weights) are published for anyone to download, run, and modify, usually under a license that sets limits on commercial use or redistribution.
- Large language model (LLM)A large language model (LLM) is a neural network trained on enormous amounts of text to predict the next word, then tuned to follow instructions and hold a conversation. LLMs power ChatGPT, Claude, Gemini, and Microsoft Copilot.
- API (application programming interface)An API (application programming interface) is a way for one piece of software to talk to another. An AI API lets your own tools, spreadsheets, or automations send a prompt to a model and receive the response, without a person typing in a chat window.
- Data privacyData privacy, in the context of AI, is the set of rules and practices that govern what information you put into an AI system, who can see it, how long the vendor keeps it, and whether it is used to train future models.
- Model weightsModel weights are the numbers inside a trained neural network that determine how it responds to input. They are what training produces and what a company ships or protects; releasing them publicly is what makes a model 'open weights.'