Bard, like ChatGPT, is a Large Language Model (LLM) trained on enormous datasets of text and code.
The biggest difference compared with ChatGPT is the amount of text it was trained on: 1.56 trillion words, while ChatGPT was trained on a dataset of "only" 175 billion words.
For now, Bard is completely free, although there is a limit on the amount of generated text. The first tests are positive in any case, and a little competition certainly cannot hurt the evolution of both.