{"id":"32837","prompt":"Goal: Create a 2-column by 3-row editorial comparison grid of six AI image model outputs, each showing the same close-up jewelry photography subject: a cropped side portrait of a punk/goth person’s ear wearing a black plug-style earring with a dangling silver cross charm.\n\nCanvas: Landscape collage, approximately 1200×1000 px, no outer border, six equal rectangular panels touching edge-to-edge in a 2×3 grid.\n\nLayout: Use exactly 6 panels with small rounded white label pills near the top-left of each panel. Panel labels, in reading order, must be: 1) MIDJOURNEY V8.2, 2) GROK IMAGINE 2.0, 3) SEEDREAM 5 PRO, 4) REVE 2.0, 5) NANO BANANA PRO, 6) GPT 2. Keep the labels compact, uppercase, black condensed sans-serif text on off-white rounded rectangles.\n\nSubject details: Every panel is an extreme close-up editorial portrait crop focused on a black circular plug earring with a dangling silver cross charm. Show only part of the face and ear: cheek texture, ear cartilage, stray dark hair strands, and sometimes a partial eye, nose ring, lips, neck, or jawline at the edge of the crop. The person has realistic pale skin with visible pores, redness, shine, blemishes, and uneven texture and dark messy hair. The cross charm should be metallic, reflective, and slightly different in shape and scale across panels, as if generated by different models.\n\nPanel variations: Top-left panel is the widest facial crop, showing the left side of the face with a partial eye, nose ring, lips, cheek, ear, and small silver cross; top-right is a tight ear crop with strong pink-red flash lighting and a bright cross; middle-left is similar to top-right but closer, with the cross blown out by flash; middle-right is a darker purple-toned side-of-head crop with the ear and smaller cross against a near-black background; bottom-left is a gritty flash photo with pronounced facial stubble/pores, black plug, and bright silver cross against a dark red-black background; bottom-right is an extreme macro of the earlobe with a black oval plug and large glowing silver cross, warm skin tones and magenta shadows.\n\nVisual style: raw editorial macro jewelry photography, harsh on-camera flash, high contrast, glossy skin highlights, shallow depth of field, gritty realism. Use a dark club-like background with black, burgundy, and purple tones. Make the collage feel like a side-by-side model comparison posted on social media, with each panel having subtly different realism, anatomy, lighting, and jewelry rendering.\n\nConstraints: Do not add any title outside the six panels. Use exactly the six label pills listed above and no extra captions. Keep the jewelry as the visual focus in every panel. Avoid clean studio beauty retouching; preserve pores, redness, flash glare, and imperfect skin detail.","promptEn":"Goal: Create a 2-column by 3-row editorial comparison grid of six AI image model outputs, each showing the same close-up jewelry photography subject: a cropped side portrait of a punk/goth person’s ear wearing a black plug-style earring with a dangling silver cross charm.\n\nCanvas: Landscape collage, approximately 1200×1000 px, no outer border, six equal rectangular panels touching edge-to-edge in a 2×3 grid.\n\nLayout: Use exactly 6 panels with small rounded white label pills near the top-left of each panel. Panel labels, in reading order, must be: 1) MIDJOURNEY V8.2, 2) GROK IMAGINE 2.0, 3) SEEDREAM 5 PRO, 4) REVE 2.0, 5) NANO BANANA PRO, 6) GPT 2. Keep the labels compact, uppercase, black condensed sans-serif text on off-white rounded rectangles.\n\nSubject details: Every panel is an extreme close-up editorial portrait crop focused on a black circular plug earring with a dangling silver cross charm. Show only part of the face and ear: cheek texture, ear cartilage, stray dark hair strands, and sometimes a partial eye, nose ring, lips, neck, or jawline at the edge of the crop. The person has realistic pale skin with visible pores, redness, shine, blemishes, and uneven texture and dark messy hair. The cross charm should be metallic, reflective, and slightly different in shape and scale across panels, as if generated by different models.\n\nPanel variations: Top-left panel is the widest facial crop, showing the left side of the face with a partial eye, nose ring, lips, cheek, ear, and small silver cross; top-right is a tight ear crop with strong pink-red flash lighting and a bright cross; middle-left is similar to top-right but closer, with the cross blown out by flash; middle-right is a darker purple-toned side-of-head crop with the ear and smaller cross against a near-black background; bottom-left is a gritty flash photo with pronounced facial stubble/pores, black plug, and bright silver cross against a dark red-black background; bottom-right is an extreme macro of the earlobe with a black oval plug and large glowing silver cross, warm skin tones and magenta shadows.\n\nVisual style: raw editorial macro jewelry photography, harsh on-camera flash, high contrast, glossy skin highlights, shallow depth of field, gritty realism. Use a dark club-like background with black, burgundy, and purple tones. Make the collage feel like a side-by-side model comparison posted on social media, with each panel having subtly different realism, anatomy, lighting, and jewelry rendering.\n\nConstraints: Do not add any title outside the six panels. Use exactly the six label pills listed above and no extra captions. Keep the jewelry as the visual focus in every panel. Avoid clean studio beauty retouching; preserve pores, redness, flash glare, and imperfect skin detail.","promptZh":"目标：创建一个 2 列 3 行的编辑类对比网格，展示 6 个 AI 图像模型的输出结果。每个网格均展示相同的珠宝摄影主题：佩戴黑色耳扩及银色十字架吊坠的朋克/哥特风格人物侧面特写。\n\n画布：横向拼贴，尺寸约为 1200×1000 像素，无外边框，6 个等大的矩形面板以 2×3 网格形式无缝拼接。\n\n布局：使用 6 个面板，每个面板左上角附近放置一个小巧的圆角白色标签。面板标签按阅读顺序依次为：1) MIDJOURNEY V8.2, 2) GROK IMAGINE 2.0, 3) SEEDREAM 5 PRO, 4) REVE 2.0, 5) NANO BANANA PRO, 6) GPT 2。标签需保持紧凑，使用黑色压缩无衬线字体，背景为米白色圆角矩形。\n\n主体细节：每个面板均为极近距离的编辑类人像裁剪，聚焦于 佩戴黑色圆形耳扩及银色十字架吊坠。仅展示部分面部和耳朵：脸颊纹理、耳廓、凌乱的深色发丝，有时包括裁剪边缘的部分眼睛、鼻环、嘴唇、颈部或下颌线。人物具有 真实的苍白皮肤，可见毛孔、红血丝、油光、瑕疵及不均匀的纹理 以及凌乱的深色头发。十字架吊坠应呈现金属质感和反光效果，且在不同面板中形状和比例略有差异，模拟不同模型的生成效果。\n\n面板差异：左上角面板为最宽的人像裁剪，展示左侧脸部，包含部分眼睛、鼻环、嘴唇、脸颊、耳朵及小巧的银色十字架；右上角为紧凑的耳朵特写，采用强烈的粉红色闪光灯照明，十字架明亮；中左侧与右上角相似但更近，十字架被闪光灯过曝；中右侧为偏暗的紫色调头部侧面裁剪，耳朵和较小的十字架置于近乎黑色的背景前；左下角为粗颗粒感的闪光灯照片，突出面部胡茬/毛孔、黑色耳扩及深红黑色背景下的明亮银色十字架；右下角为耳垂的极度微距，配有黑色椭圆形耳扩和巨大的发光银色十字架，肤色偏暖，阴影呈洋红色。\n\n视觉风格：原始编辑类微距珠宝摄影，强烈的机顶闪光灯，高对比度，光泽感皮肤高光，浅景深，粗颗粒写实感。使用类似夜店的深色背景，包含黑色、酒红色和紫色调。使拼贴画呈现出社交媒体上常见的模型对比效果，每个面板在写实度、解剖结构、光影和珠宝渲染上均有细微差别。\n\n约束：不要在 6 个面板之外添加任何标题。仅使用上述 6 个标签，不要添加额外说明文字。确保珠宝在每个面板中均为视觉焦点。避免过度修饰的影棚美颜效果；保留毛孔、红血丝、闪光灯眩光及不完美的皮肤细节。"}