# Open source vs open weights

Concept · [Building with AI](https://ecosystem.samcarlton.com/ai)

Also called: open weights, open-weight, open weight model, open-weights model, open model, open models, open source AI, open source model, closed model, closed models, proprietary model, closed source model, closed weights

Two labels for [AI models](https://ecosystem.samcarlton.com/model) that people often mix up. Open weights means the [trained](https://ecosystem.samcarlton.com/training) numbers of a model are published, so anyone can download and run it. Open source AI, as the [Open Source](https://ecosystem.samcarlton.com/open-source) Initiative defines it, asks for more: [the weights](https://ecosystem.samcarlton.com/weights), the code used to train and run the model, and detailed information about the training data, on terms that let anyone use, study, change and share the system. A closed or proprietary model publishes none of these, and people reach it only through its maker's app or [API](https://ecosystem.samcarlton.com/terms/api).

**For example:** A lab [releases](https://ecosystem.samcarlton.com/deployment) a model's [weights](https://ecosystem.samcarlton.com/weights) but not its training code or a description of its data. Under the [Open Source](https://ecosystem.samcarlton.com/open-source) Initiative's definition that model is open weights and is not open source AI. A team that wants to run a model on its own [hardware](https://ecosystem.samcarlton.com/hard-tech) needs at least the weights.

## Connects to (12)

- [Weights](https://ecosystem.samcarlton.com/weights): The numbers inside a model that it learned during training.
- [Open source](https://ecosystem.samcarlton.com/open-source): Software whose code is published for anyone to read, use and build on, under a license that sets the rules for doing so.
- [Local model](https://ecosystem.samcarlton.com/local-model): A model that is downloaded and run on a person's own computer or a company's own servers, not on a provider's.
- [Model](https://ecosystem.samcarlton.com/model): The trained system at the center of an AI product.
- [Hugging Face](https://ecosystem.samcarlton.com/hugging-face): A platform where the machine learning community shares AI models, datasets and demo apps.
- [Frontier model](https://ecosystem.samcarlton.com/frontier-model): One of the most capable general-purpose AI models available at a given time.
- [LoRA](https://ecosystem.samcarlton.com/lora): Low-rank adaptation.
- [Llama](https://ecosystem.samcarlton.com/llama): Meta's family of AI models whose weights are published for download.
- [Model Garden](https://ecosystem.samcarlton.com/model-garden): A catalog in Google Cloud's Gemini Enterprise Agent Platform, formerly Vertex AI, where teams find and deploy AI models from Google and…
- [Gemma](https://ecosystem.samcarlton.com/gemma): Google's family of open-weights AI models, small enough that some run on a laptop or phone.
- [LM Studio](https://ecosystem.samcarlton.com/lm-studio): A free desktop app for downloading open AI models and running them on a person's own computer, with a chat window and a local API.
- [Google DeepMind](https://ecosystem.samcarlton.com/google-deepmind): Google's AI research team, formed by combining Google Brain and DeepMind.

Map view: https://ecosystem.samcarlton.com/open-source-vs-open-weights