Original document
GPT Image 2 · en
A surreal, photorealistic night scene: a vast, calm, mirror-like expanse of water against a deep, star-studded sky. A bright full moon hangs in the upper left, with the Milky Way and faint blue-green aurora extending into the background. In the real world, everything remains simple and realistic: a slender woman with shoulder-length dark hair walks barefoot across the water from left to right, wearing a {argument name="dress color" default="black"} She wore an elegant, slit evening gown. Her posture was natural and composed, her face softly blurred or out of focus. The water was smooth as a mirror, with subtle ripples and a realistic moonlight reflection. The symbolic magic was primarily placed in the reflection: at her feet, instead of a normal reflection, a luminous, ethereal blue-white line art tiger appeared in the water, its body curled up, shimmering like energy or moonlight. The tiger was sniffing or approaching a luminous rose near the lower right corner of the reflection. Around the rose, six tiny, rising, luminous butterfly-shaped points of light were added, forming soft arcs. The work employed cinematic lighting, high contrast, cool blue tones, a dreamlike atmosphere, ultra-refined reflection effects, and a clean, artistic fantasy photography style. A delicate cursive title was added at the center of the bottom edge of the image, containing the following text:{argument name="caption text" default="In me the tiger sniff the rose."} The composition uses rounded corners to present the aesthetic appeal of a vertical social media screenshot, but the main artwork should be the visual focal point.How to use
- Copy the complete original prompt, including any parameters.
- Replace the subject, text and composition with your own requirements.
- Use a compatible external tool; check its current access conditions and review the result.
Before you start
- Check whether a reference image is required. A compatible model label does not mean this example was generated with that model.
- Keep names, figures and visible text accurate; results can vary between model versions.
