WHAT NEW IN X11.01 SUN

X 上的 AI 合集档案 · 22 个合集 · 64,204 条

gpt-imageIMAGE

GPT Image 2.5 城市微缩模型提示词模板

一条可直接套用的 GPT Image 2.5 提示词:把任意城市做成印度漆木玩具风格的 4:5 电影感微缩立体模型。

Saul Goodman
@Goodmanprotocol2026.09.19收录原推发布于 2026.09.18
GPT Image 2.5 城市微缩模型提示词模板(第 1 张,共 4 张)
GPT Image 2.5 城市微缩模型提示词模板(第 2 张,共 4 张)
GPT Image 2.5 城市微缩模型提示词模板(第 3 张,共 4 张)
GPT Image 2.5 城市微缩模型提示词模板(第 4 张,共 4 张)
Saul Goodman @Goodmanprotocol原推发布于 2026.09.18
在 X 上打开原推

A whole city, carved into a moment. GPT Image 2.5 Prompt: [LOCATION] - Premium 4:5 Cinematic Travel Diorama Create a premium 4:5 cinematic travel diorama of [LOCATION], designed as a handcrafted Indian lacquered wooden-toy world. Everything should feel physically carved, painted and collectible, with glossy woodgrain, rounded handcrafted forms, intricate folk-art detailing, warm golden-hour illumination, soft cinematic bokeh and shallow depth of field. Composition: Place one lone toy traveler in the lower-left foreground, photographing the destination. A single iconic landmark or hero object dominates the middle-right. Add only 2–3 unmistakable location-specific environmental cues, while secondary architecture or landscapes softly recede into the background. Use a textured wooden or stone foreground to create depth. Transform all elements of the destination—architecture, terrain, vehicles and props—into the same cohesive handcrafted miniature style. Maintain generous negative space, strong visual hierarchy, warm color harmony, tactile craftsmanship and elegant cinematic storytelling. Use only the most recognizable elements of [LOCATION]. Avoid crowds, clutter, excessive landmarks, generic tourist imagery, text, logos, poster-style layouts and photorealistic elements. The final image must feel like one unified handcrafted miniature world, not toys placed over a background.

x.com/goodmanprotocol/status/2100806204520034392采集于 2026.09.19 10:07