A. Smyntyna / Code
ENRU

open weights

A model whose trained parameters are published as files you can download and run. Not the same as open source: the training data and the code that produced the weights are usually withheld. Llama, Qwen, Gemma, DeepSeek and gpt-oss are open weights. Fable and GPT are not.

What you get is the finished artifact, under a license that ranges from MIT to one with a user cap and a list of forbidden uses. Read it before shipping anything.

What open weights actually buy the rest of us is a reference point. Nobody publishes the size of a closed frontier model, so the only way to reason about what one would cost to run is to look at the largest open model of the same generation and do the arithmetic on that.

Read next