When using the ChatGPT image generation model, I noticed that previously generated images appeared together in a single frame, which made comparison easier. Now, they are in separate frames, complicating the workflow. Restoring the previous layout would make reviewing image quality more convenient.