Google

A motion-injection edit test. The model adds wind-driven movement (blowing hair, flying leaves) to a still photo while leaving the rest of the scene intact. The difficulty is making motion look physically plausible without distorting the subject or background.

A style-transfer test. The model must convert a photo into a convincing hand-painted anime look while preserving the original composition and subject, so the result reads as a genuine illustration rather than a filter laid over the photo.

A stylized edit test. The model has to exaggerate features for a caricature while keeping the subject recognizable, and work in the requested story elements (profession, dogs, hockey). It balances recognizable likeness against deliberate distortion.

A targeted edit test. Adding a realistic head of hair requires a believable hairline and texture while keeping the face, expression, and lighting unchanged. Common failures are wig-like or pasted-on hair and subtle changes to the underlying face.

Combines two references (a person and a car) into one believable driving scene. It checks whether the model keeps the man’s identity and the car’s appearance while placing them convincingly on the coastline with matched lighting and perspective.

This prompt is challenging because it forces models to do two hard things at once: accurately transfer complex body poses and joint angles from Image 1 while strictly preserving the exact face, identity, and clothing style from Image 2. Most models struggle with unnatural limb distortions, identity blending, incorrect fabric behavior on the new pose, and mismatched lighting, making it a strong test of advanced spatial reasoning and character consistency.

This is extremely practical (think e-commerce, fashion, virtual dressing rooms). It’s visually very obvious when it fails, and success looks impressive. It tests a model’s understanding of clothing physics, body shape, and lighting.