Some AI applications are freely available. Examples include language models from French vendor Mistral or the Llama family from Facebook/Meta. However, it is not correct to call these “open source”. What you get as a user is the end result of the training, the core of a large language model called “weights”. At the same time, the training material or the training process is not automatically included.
In other words, such AI models can be used for free and even modified and refined within certain limits (“fine tuning”). But you don’t get all the building blocks you need to create your own AI from scratch.
For this reason, the term “open weights” is often preferred to indicate this important difference. Unfortunately, there is no generally accepted definition so far.