Figcaptions versus alt text

DRANK

How does an image’s descriptive (alt) text differ from the caption it would have if it were used in a <figure> element?