What is the best model for this prompt?

This prompt is optimized for Sora 2 but can be adapted for other advanced image generators that support reference images and style control.

Can I change the subject or setting?

Yes, you can modify the gender, clothing, or bar type, but maintain the core instructions for pose and candid flash style for best results.

Will the face be anonymized?

Yes, the prompt includes instructions to keep facial anonymization, which the AI should preserve unless you explicitly remove that part.

Izakaya Beer Drinking Photo Variant — AI Prompt | sora2hub

GPT Image 2AI 提示词

Izakaya Drinking Variant — Sora 2 Photo Prompt

Create a realistic photo variant of a woman in an izakaya, adjusting her pose to raise a beer glass while preserving the candid flash and atmosphere.

提示词 · EN

使用提供的参考图，在居酒屋场景内创作一张主体相同且逼真的照片变体。保持女性坐在凳子上的姿势、穿着和整体位置不变，但调整她的姿势，使其稍微转向镜头和身旁的男性，并举起一个小啤酒杯送到嘴边，做出饮酒的动作。添加装有金色啤酒和泡沫的啤酒杯，并保持与参考图一致的抓拍闪光灯效果、昏暗的酒吧背景以及自然的拥挤氛围。保留对脸部的匿名化遮挡处理。

关于这个提示词

Create a realistic photo variant of a woman in an izakaya, adjusting her pose to raise a beer glass while preserving the candid flash and atmosphere. 它更适合作为 GPT Image 2 的概念艺术起点：先保留画面结构、主体关系和风格约束，再替换成你的品牌、人物或场景。

使用时建议先小幅修改主体、场景、镜头和色调，再生成多个版本对比构图与细节。这样页面内容对用户有实际帮助，也避免把模型名或标签机械堆在正文里。