Lesson 40 · 2 min
Open vs closed models
Some AI models can be downloaded and run by anyone, while others are only available through the company's app or service.
A closed model is kept by the company that made it. You use it through their website, app or API. You cannot see or download its weights, the learned numbers that make it work.
An open model, often called open-weight, has its weights published. Anyone can download it, run it on their own computer or server, and adjust it. Licences differ, so some allow any use and some add limits.
Closed models are often the most capable and the easiest to use. Open models give more control: your data can stay on your own machine, you can fine-tune them, and you are not tied to one company.
Many teams use both: a top closed model for hard tasks and an open local model for private or high-volume work.
Do it now
Sort the models you know
- 1. List the AI tools you use or have heard of.
- 2. Ask a chatbot the prompt below to sort them.
- 3. Pick one task where keeping data on your own machine would matter to you.
For each of these AI models or tools, tell me whether it is open-weight or closed, and what that means for privacy and cost: [LIST]. If you are unsure about one, say so.
Quick check
Words from this lesson
- Open-source model
- A model whose trained weights are published so anyone can download, run and adapt it, often on their own hardware.
- Weights
- The learned numbers that make up a trained model. Sharing a model usually means sharing its weights file.
- Local model
- An AI model that runs on your own computer or phone, so your data does not leave the device.