GPT Image 2.5 prompt: cinematic night-street fashion portrait
A GPT Image 2.5 prompt for a photorealistic night-street fashion editorial portrait with warm orange and cyan lighting, motion blur and film grain.
GPT image 2.5 image Prompt 👇 A cinematic, photorealistic fashion editorial portrait of a young woman standing on a city street at night, with a short, voluminous dark brown wavy bob haircut, soft curtain bangs, natural glowing skin, subtle freckles, warm blush, glossy nude lips, and minimal natural makeup. She wears an oversized white crewneck sweatshirt with relaxed, voluminous sleeves and high-waisted blue denim jeans, with both hands casually tucked into her front pockets. She is positioned in the center of the frame, photographed from the hips up, looking directly into the camera with a calm, confident, effortlessly cool expression. The background features a dark urban street with dramatic orange-red streetlights, vibrant cyan and teal illumination, and horizontal streaks of red, orange, and turquoise light created by long-exposure motion blur from passing traffic. Lighting and color grading: Warm orange-red backlighting illuminating the background, cool cyan fill light along the lower sides, direct on-camera flash softly illuminating the subject's face and white sweatshirt, strong contrast between warm and cool tones, deep cinematic shadows, subtle film grain, slightly dreamy analog-film texture. Photography style: Contemporary streetwear campaign, high-end fashion photography, 35mm film aesthetic, candid editorial mood, realistic skin texture, sharp facial focus, shallow depth of field, motion-blurred background, natural proportions, imperfect film-like details, balanced composition, authentic nighttime city atmosphere. Vertical portrait, 4:5 aspect ratio, medium shot, subject centered, ultra-realistic photography, high detail, premium fashion magazine quality. Avoid: cartoon rendering, CGI, plastic skin, excessive retouching, oversaturated colors, distorted anatomy, extra fingers, blurry face, artificial facial features, text, logos, watermarks.