1. It couldn't recognize color and couldn't understand the contextual color information. When I requested "white background", It keeps giving me blue background. 2. For different categories and different inputs, it gives me exactly the same images. For example, when I switch from logo to book cover, or to other categories, and change a little bit of my inputs, it keeps giving me the same images. 3. It erased what I just typed in the “input”, and it has happened several of the times, which largely wasted my time and it’s very annoying. 4. Besides color, it’s hard to understand my other requests as well. Like number, when I input “one blueberry “, it keeps showing a bunch of blueberries. The contextual understanding is really not expected.