I don't think that's a fair comparison because they're fulfilling substantially different niches. Gemini is a conversational model that can generate images, but is mainly designed for text. Stable Diffusion is only for images. If you compare a model that can do many things and a model that can only do images by how well they generate images, of course the image generation model looks better.
Stability does have an LLM, but it's not provided in a unified framework like Gemini is.
No comments yet.