GPT-Image 1.5 vs GPT-Image 2

Prompt

"Palm trees and pastel art-deco buildings blur behind a tiny roadside farm stand set right on a Miami promenade, with a handwritten price board, stacked mango crates, and a glimpse of a small pen of animals in the background. Tight collarbone-up framing on a mixed-race woman in her mid-20s with locs, sun-sheened skin, and big expressive anime eyes narrowed in visible frustration—jaw clenched, brows pinched, one hand gripping her phone while the other tugs her tank top strap, sweat beading at her temple and a chunky gold hoop earring catching the light. A floppy-eared rabbit naturally edges into the lower corner of the frame near the crate rim, ears drooping as it sniffs at a fallen strawberry; cel-shaded modern anime key visual, vibrant color pops, shallow depth of field, intimate social-media candid vibe."

Prompt

"Golden hour spills between NYC brownstones as a friend snaps a candid mid-stride on the sidewalk—anime cel-shaded key visual with warm lens flare blooming at the frame edges and backlight haloing everything. A mixed-race guy in his mid-20s with straight shoulder-length hair walks calmly with relaxed shoulders and an easy half-smile, one hand in his pocket and the other loosely holding an iced coffee; he wears a thrifted bomber over a plain tee, tapered chinos, clean low-top sneakers, and a canvas tote slung across his chest. Behind him: stoop railings and stone steps, a leafy street tree casting dappled shadows, a parked yellow cab and a bike locked to a signpost, a corner bodega awning in soft blur—peaceful, unhurried body language and serene eyes catching the sun."