Open-Weight Model
An AI model whose trained parameters are publicly downloadable, so anyone can run or modify it themselves.
01 Definition
An open-weight model publishes its trained parameters for anyone to download and run on their own hardware or cloud servers, as opposed to a closed model you can only access through a company's paid API or app.
02 Why It Matters
Open-weight models let developers self-host for privacy, customize via fine-tuning, and avoid per-request API costs — a real alternative to always renting access to a closed model.
03 Example
Several of Mistral AI's models are released with open weights, while GPT-4-class models are closed.