
Original document
GPT Image 2 · zh
A delicate anime-style night scene illustration, using a 1:2 vertical composition. A young Japanese working woman stands on a riverside promenade at night, facing the viewer, cheerfully raising a large glass of beer in a toast. She wears a fitted dark navy blue business suit, paired with a pencil skirt, white shirt, and red tie, with her company badge around her neck. She has long, straight, dark brown to black hair with soft bangs, and a slender figure, exuding a relaxed and friendly atmosphere after get off work. Due to the perspective, her right hand, which is extended towards the camera, holds an ice-cold beer glass filled with golden lager and foam, making it particularly eye-catching in the foreground. Her left arm hangs naturally at her side. A white, manga-style speech bubble is added to her upper right corner, containing the vertically arranged Japanese text "今日もお疲れさまでした!" (I'm tired today!). The scene is set on a paved waterfront promenade with metal railings, three visible warm-colored lanterns beside it, and a river or canal in the background against a modern city skyline with many bright windows. Above, a clear, deep blue starry sky dominates the space, with the Milky Way extending diagonally from below the center behind the character to the upper right corner, forming a bright band of light. Cinematic lighting effects are used, with cool-toned moonlight contrasting with warm-toned lantern light, clean lines, delicate sheen and shadow on the suits, and rich detail in the starry sky, creating a tranquil yet celebratory after-get off work atmosphere.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.





