open-weight vs closed LLM

Some LLMs are open-weight: the company publishes the trained parameters so anyone can download them and run the model on their own computers, inspect it, or modify it. Others are closed: you can only use them through the company's website or an API, and the actual weights stay private. It is roughly the difference between owning a copy and renting access.

The tradeoff is control versus convenience. Open weights give privacy, offline use, and the freedom to fine-tune, but you supply the hardware and expertise. Closed models are easy to call and often the most capable, yet you depend on the provider's pricing, availability, and rules. Note that open-weight rarely means fully open: the training data and code are usually still secret.