Inpaint images using prompts
Generate highly aesthetic images
Request evaluation for new speech models