Often, when asked about the capabilities of models, GPT-4 appears impartial while Bard insists on saying that it is better at “answering open ended, challenging, or strange questions”. Apparently Bard was trained to answer that.
It’s funny to see Google’s approach.