Hey rameez, this is an issue with the underlying image models. The image models and text models are different. The text models produce text. The image models produce images but do not have an understanding of language. So they can’t spell, unfortunately. They recognize words as images not as symbolic systems.