动漫人物 Text Effect
用原作的模型 OpenAI GPT-Image-2.5 Sunburst · 15 积分
这是一组配对头像,绘制成一幅没有标题文字的宽幅动漫风插画。蓝光下的银发少年和红光下的黑发少女都戴着发光耳机,穿着黑色连帽外套,身后是霓虹城市的高楼。你可以更换发色和瞳色、耳机光效、服装和背景,做成属于自己的配对头像。
Create one wide 2:1 image designed as two matching square profile pictures side by side. The vertical center is the boundary between the two equal square halves. Each half must work alone as a complete profile picture and remain clear when circle-cropped: keep each character's entire face, hair silhouette, and headset comfortably inside the central 78% of their own square. No face, head, or body crosses the center boundary. Show two original anime gamer friends standing back to back, shoulders touching near the center seam, with their bodies angled inward but their faces turned toward the viewer. Left square: one character in a vivid cobalt-blue gaming headset and blue-accented streetwear. Right square: the other character in a vivid crimson-red headset and red-accented streetwear. Give them distinct original hairstyles and expressive, friendly-confident expressions. Head-and-shoulders framing, large layered irises with two bright catchlights per eye, small simple noses and mouths, subtle blush. Clean anime line art, heavier on the outer silhouettes, crisp cel shading with one hard shadow tone and a thin colored rim light. Their pose and color accents should answer each other as a deliberate matching pair. Behind them, a softly blurred rainy neon city at night, shared across the center seam: deep navy and charcoal buildings, small cyan and magenta signs rendered as abstract light shapes only, luminous rain bokeh, a few fine rain streaks, subtle reflections. Blue light rims the left character; red light rims the right, with a faint violet glow where the colors meet. Keep the city low-detail and subdued so the faces dominate at tiny profile-picture size. Bold, simple shapes, strong face-to-background contrast, clean separation between both portraits. No text, no letters, no logos, no interface, no extra people. Finished illustration fills the canvas edge to edge.
作品标签
生成于 2026年10月23日 · 1456 × 720 · JPEG · OpenAI GPT-Image-2.5 Sunburst · 版式:Twin Portrait Split · 公开作品 · 可免费商用(CC BY 4.0)

















