Here is the the list

• 1024 x 1024 (1:1 Square)
• 1152 x 896 (9:7)
• 896 x 1152 (7:9)
• 1216 x 832 (19:13)
• 832 x 1216 (13:19)
• 1344 x 768 (7:4 Horizontal)
• 768 x 1344 (4:7 Vertical)
• 1536 x 640 (12:5 Horizontal)
• 640 x 1536 (5:12 Vertical, the closest to the iPhone resolution)

My favorite size is 1344×768 because  when you upscale it with 1.43 ratio you get 1920×1096 perfect for web and devices.

SDXL trained on 1024 x 1024 size but fine-tuned on this list of sizes.

Detailed Explanation about SDXL sizes and where to use each size

When creating images with Stable Diffusion, one important consideration is the image size or resolution. The AI model was trained on images of varying sizes, so you can generate results at different resolutions. Here are some of the most common and useful image sizes to use with Stable Diffusion:

• 1024 x 1024 (1:1 Square) – The default size that many people use. It provides a nice balance of detail and generation speed.
• 1152 x 896 (9:7) – A common photo print size. Good for portraits.
• 896 x 1152 (7:9) – The inverse of the 9:7 photo print size. Also good for portraits.
• 1216 x 832 (19:13) – A widescreen horizontal rectangle. Useful for landscapes and wide images.
• 832 x 1216 (13:19) – The vertical version of the 19:13 widescreen size.
• 1344 x 768 (7:4 Horizontal) – A standard widescreen computer/laptop resolution. More horizontal space to work with.
• 768 x 1344 (4:7 Vertical) – The vertical equivalent of the 7:4 horizontal size.
• 1536 x 640 (12:5 Horizontal) – Very wide horizontal rectangle. Can be great for extra wide images.
• 640 x 1536 (5:12 Vertical) – Very tall vertical rectangle. Useful for portraits and long vertical slices. Close to iPhone resolution.

My personal favorite is 1344×768 because when you upscale it by 1.43x, you get 1920×1080 which is perfect for web and devices. The 7:4 aspect ratio provides a nice wide canvas to work with too.

Experimenting with different image sizes and ratios can lead to all kinds of new creative possibilities with Stable Diffusion. Just keep in mind that larger sizes will take longer to generate. Overall, the size you choose depends on your specific needs and the type of images you want to create.