SnowPro Specialty: Gen AI free question 4 / 10 · Domain 2: Snowflake Gen AI Functions
A quality team wants a model to compare a reference photo of a part with a photo from the production line, both in a stage, and describe the differences in one answer. What does the multimodal COMPLETE page say?
- Images are read from internal stages; external ones are not
- Images are first turned into text, then compared as text
- Several images can be processed in one operation, e.g. comparison
- Each call takes one image, so the two answers are joined later
Show the answer and explanation
Answer: C
Why this is correct
The multimodal COMPLETE page says that, given an image and a prompt, it generates a response using a language model, supports image models along with text models, and processes images stored in an internal or external stage. It can be used for a single image, for multiple images in a batch with the same or a different prompt for each, or for multiple images in a single operation, for example comparison.
Why the other options are wrong
"Each call takes one image, so the two answers are joined later" … Single images are one supported pattern, not the only one. The page lists processing several images in one operation.
"Images are first turned into text, then compared as text" … The function takes an image and a prompt and generates a response with a model that supports images. No conversion to text is needed.
"Images are read from internal stages; external ones are not" … The page says the function processes images stored in an internal Snowflake stage or an external stage.
Which surface to use
The page is kept for backward compatibility and points new work to AI_COMPLETE. The same idea, an image passed as a file reference together with a prompt, carries over.
Go to the SnowPro Specialty: Gen AI practice app (10 free questions; purchase unlocks all questions)