The company is working on an image and video model and new text-based model, which aids with coding.