ChatGPT Images 2.6 Prompt Library
18 works made with ChatGPT Images 2.6 (GPT-Image-2.6) by creators on X, each with 5,000+ views on the original post, GIFs playable here. 5 include the prompt the creator shared, kept verbatim with its source. Updated daily.
Updated daily · last update 2026-11-05 · view counts as of 2026-11-05
请以我上传的角色图为参考,按「表情包主题」做一张可下载的动态 GIF。 【角色保持一致】锁发型/配色/标志道具,Q 版比例全帧不变 【自动设计动作】一个核心笑点,夸张一点,首尾能循环 【自动配字】2~8 个字网感短句,顶部粗圆体+白描边 【画面】正方形纯白底,2~3 秒循环,聊天表情包规格 表情包主题:【填你的梗,比如「周一不想上班」】
create a GIF of a man doing a standing backflip
请以这个内容创建一个 GIF,内容是: # JAPANESE SUMMER|夏日元气少女 **10秒日系青春活力MV视频提示词** **画面规格:** 9:16 竖屏,10秒,30fps,写实摄影质感。 **人物设定:** 22岁成年东亚女性,清透冷白皮肤,短款深棕色波波头,轻薄空气刘海,五官甜美可爱,身材匀称。穿着参考图中的蓝色泳装、黄色镂空针织罩衫及牛仔短裤。人物保持统一,性格活泼俏皮,笑容自然,有青春少女的灵动感。 **整体风格:** Japanese Youth MV / Japanese…
请以我上传的大肥鱼图片为角色和画风参考,根据最后一行的「表情包主题」,直接制作一张可下载的动态 GIF 表情包。 【角色保持一致】 保留参考图中的蓝色渐变长发、呆毛、鲸鱼耳鳍、蓝色蝴蝶结、白色女仆头饰、蓝白女仆裙和大鲸鱼尾巴。 保持原图的大头小身体 Q 版比例、圆脸和可爱画风,人物外观在所有帧中保持一致。 【自动设计动作】 理解我想表达的情绪或梗,将它设计成一个简单、有趣、一眼能看懂的短动作。 突出一个核心笑点,用夸张表情、动作反差和轻微的压缩回弹增加喜感。头发、呆毛、耳鳍和尾巴自然跟随动作。…
使用这个角色,生成一张GIF图,内容是她吃大白米饭,图片里有文字:真香!
FAQ
What is ChatGPT Images 2.6?
OpenAI's image model released on November 4, 2026, available to every ChatGPT and Codex user. It adds animated GIFs (2 to 16 frames) and saved reference images for people, pets and characters. In the API it is GPT-Image-2.6, in two tiers: Flare (faster) and Sunburst (higher quality).
Does every work come with a prompt?
No. 5 of the 18 works include a prompt. Prompts come only from the creator: the post itself, their own reply, a screenshot in a reply, or a link they shared, kept verbatim. Works without a published prompt are listed too and link straight to the original post; tick "Only with prompts" to filter.
What gets included?
The original post has 5,000+ views, the images, GIF or video are attached to that post, and the creator says it was made with ChatGPT Images 2.6 (GPT-Image-2.6). Reposts, official demos and platform promotions are left out. Model attribution is as stated by the creator, not independently reproduced. New works are added daily.
What does "Needs your own image" mean?
Besides the text prompt, the creator gave the model their own photo, sketch or a reference image. Swap in your own image to get a similar result.
I'm a creator and want my credit fixed or a work removed.
Send a DM to @GoSailGlobal on X and it will be handled quickly. All works belong to their creators; listing them here does not imply any license.
Can I use the data?
Yes. Structured data for every work (post link, creator, category, views, prompt source) is open on GitHub at https://github.com/zhuyansen/awesome-chatgpt-images-2.6 and synced daily.
See also: Claude 5.5 Prompt Library (video, motion, 3D, games) · GitHub