Images, upgraded—introducing ChatGPT Images 2.6. Now with GIFs and reference images. Plus, two new models are coming to the API with 50% faster image generation.
ChatGPT Images 2.6 AI 生成器
即将推出使用 ChatGPT Images 2.6 创作插画、制作海报或编辑照片。即使构图已经完成,也可以调整标题或光线,同时保留让角色、产品或品牌易于辨认的细节。
ChatGPT Images 系列使用指南
看看如何用文字创作说明和参考照片引导图像创作,从初稿一路修改到第二轮。
ChatGPT Images 2.6 AI 图像模型的核心功能
看看以 LIFE 字样为主题的海报,以及股价下跌和上涨时的两段反应动画。再看看溪边的水獭静态图和像素风灯塔如何变成循环短动画。
字母拼出的人生与两种股市反应

Create an ultra-photorealistic, cinematic 3D advertising poster in a perfect 1:1 square aspect ratio, rendered in stunning 8K quality. Main Concept — Four Stages of Life Depict four different age versions of the same beautiful European-American female, aged 7, 14, 21, and 26, arranged from left to right in ascending age and physical size. The woman has naturally beautiful Western facial features: fair luminous skin, captivating blue-gray eyes, long softly wavy chestnut-brown hair, a refined straight nose, delicate rosy lips, elegant facial proportions, and a warm, charming smile. All four versions must share highly consistent facial identity, eye shape, nose structure, hair color, skin tone, and distinctive features. Age progression should be biologically realistic, making them instantly recognizable as the same person throughout different stages of her life. Four Age Versions and Outfits Age 7 — Playful Childhood An adorable, cheerful little girl with round youthful cheeks, bright sparkling eyes, and shoulder-length wavy chestnut hair. Wearing a cute yellow school uniform, white socks, and small sneakers. Standing playfully beside the first giant red 3D letter, gently leaning against it with an innocent, joyful expression. Age 14 — Confident Teenage Years A charming teenage schoolgirl with naturally developing facial features and long chestnut hair in a casual ponytail. Wearing a neat white school uniform, a pleated skirt, white sneakers, and a school backpack. Sitting casually in front of the second giant red letter, smiling confidently with a relaxed, youthful posture. Age 21 — Stylish Young Adulthood A beautiful young adult woman with elegant facial features, expressive blue-gray eyes, and long flowing chestnut waves. Wearing a fashionable plaid shirt, fitted blue jeans, and stylish casual sneakers. Standing confidently with her arms crossed, casually leaning against the third oversized red 3D letter. Natural fashion-model confidence and realistic body proportions. Age 26 — Elegant Adulthood The most prominent, visually striking, and slightly larger version. A stunning 26-year-old European-American woman with mature yet youthful facial beauty, luminous natural skin, and voluminous wavy chestnut hair. Wearing a fashionable red, black, and white color-block jacket over a fitted black T-shirt, stylish blue jeans, and clean white sneakers. Sitting stylishly on top of the final giant red 3D letter, one leg hanging casually downward and the other naturally bent. Relaxed, elegant posture, confident smile, looking directly toward the camera. Composition and Giant 3D Typography Arrange all four female characters from left to right, progressively increasing in height and age. Place enormous, bold, glossy, sculptural red 3D letters across the lower portion of the composition, spelling "LIFE". Each age version interacts naturally with one corresponding letter. The giant letters have a premium polished lacquer finish, realistic reflections, dimensional depth, and subtle highlights. Make the composition feel like a luxury commercial campaign celebrating a woman's journey through life. Age Labels and Graphic Details Add elegant, dark burgundy handwritten calligraphy above each respective character: - Age 7 Years - Age 14 Years - Age 21 Years - Age 26 Years Connect each label to its corresponding character with a thin, graceful, hand-drawn curved arrow. Position the labels carefully in the upper negative space without covering faces or overlapping with other labels. Background and Lighting Pure white seamless professional photography studio. Clean white cyclorama background, polished slightly reflective white studio floor, realistic contact shadows under all four characters and the oversized 3D letters. Premium cinematic studio lighting, large softboxes, realistic global illumination, subtle rim lighting, soft natural highlights, gentle shadows, and realistic reflections. Highly detailed skin pores, natural facial expressions, individual hair strands, fine fabric textures, realistic clothing folds, accurate human anatomy, and detailed footwear. Final Rendering Quality Ultra-photorealistic 3D rendering, cinematic commercial photography, high-end luxury advertising aesthetic, sophisticated composition, physically based materials, ray-traced reflections, realistic subsurface skin scattering, beautiful color grading, crisp focus, extraordinary texture detail. Strict Requirements Perfect square composition, 1:1 aspect ratio. Exactly four female characters. All four fully visible and clearly separated. Consistent facial identity throughout all ages. Realistic anatomy, naturally proportioned bodies, no duplicated characters, no extra limbs, no distorted hands, no cropped heads or feet, no overlapping faces, no cartoonish rendering, no watermark, no logos, no additional text beyond the specified typography.

Using ChatGPT Images 2.6 native GIF generation, create a downloadable auto-looping multi-frame GIF. A super-cute chibi adult office worker with an oversized round head, tiny body, messy fluffy hair, white shirt and loosened tie sits at a desk in a cozy miniature office, fixed front three-quarter camera, square near-shot. The computer monitor shows a clearly plunging stock-market line chart. Loop: hopeful look at the screen, notices the drop, eyes widen, face crumples, large cartoon tears stream down both cheeks, quietly wipes tears while still staring at the falling chart, then returns smoothly to the starting pose. Keep the same character, desk, monitor, outfit, and background. Warm pastel lighting, polished kawaii chibi 3D toy-like style, funny and sympathetic mood. No captions, words, numbers, logos, speech bubbles, or borders. Loop about two to three seconds. Output an animated GIF file only; do not output a sprite sheet, collage, or split still frames.

Using ChatGPT Images 2.6 native GIF generation, create a downloadable auto-looping multi-frame GIF. A super-cute chibi adult office worker with an oversized round head, tiny body, big sparkling eyes, messy fluffy hair, white shirt and loosened tie sits at a desk in a cozy miniature office, fixed front three-quarter camera, square near-shot. The computer monitor shows a clearly soaring stock-market line chart with a large upward arrow. Loop: notices the rise, eyes widen with delight, jumps up cheering, grabs a handful of generic paper banknotes with no readable currency markings, happily tosses them into the air like confetti while laughing, then returns smoothly to the starting pose. Keep the same character, desk, monitor, outfit, and background. Bright celebratory pastel lighting, polished kawaii chibi 3D toy-like style, funny joyful mood. No captions, words, numbers, logos, real currency symbols, speech bubbles, or borders. Loop about two to three seconds. Output an animated GIF file only; do not output a sprite sheet, collage, or split still frames.
水獭擦去一滴眼泪

Photorealistic cute small-clawed otter sitting on a mossy rock beside a clear stream, round dark eyes, wet brown fur with a cream chin, natural soft daylight, shallow depth of field, 16:9 landscape near-shot, no text, no logos, no watermark.

Using ChatGPT Images 2.6 native GIF generation, based on the previous still image, create a downloadable auto-looping multi-frame GIF. Keep the same otter, mossy stream rock, 16:9 framing, and lighting. Loop: eyes well up, one or two tears slide down the cheek, one small paw gently rubs an eye, then return to the starting pose. Fixed camera, stable background, no captions. Loop about two to three seconds. Output an animated GIF file only; do not output a sprite sheet, collage, or split still frames.
像素风灯塔的日落

Pixel-art coastal cliff sunset: about 16 colors, retro pixel style, 16:9 landscape. A white lighthouse on a rocky cliff, distant mountain silhouettes, calm sea with sun reflection, sun just above the horizon ready to set. No characters, no text, no UI, no logos.

Using ChatGPT Images 2.6 native GIF generation, based on the previous still image, create a downloadable auto-looping multi-frame GIF. Keep the same pixel-art coastal cliff, lighthouse, palette skeleton, and fixed 16:9 camera. Loop: the sun slowly sets below the horizon as sky and sea darken into a brief night, then the same sun rises again and the scene warms back to the opening brightness, returning smoothly to the start. Clear continuous sun path, stable pixels without jitter or morphing, no characters, captions, text, or UI. Loop about two to three seconds. Output an animated GIF file only; do not output a sprite sheet, collage, or split still frames.
ChatGPT Images 2.6 AI 图像生成器的优势
通过易于掌控的逐步修改完善创意,让创作方向更清晰,也不必轻易推倒有潜力的构图从头再来。
更快绘制海报与营销活动草稿
先把标题放在主体上方,再试试放在旁边。Flare 适合快速调整版式,方便你在完善作品前比较不同布局的阅读效果。
用自然语言进行高保真图像编辑
将 ChatGPT Images 2.6 用作 AI 图像编辑器,更换背景、调整色彩或修改物体。明确指出要保留的内容,让编辑结果贴近原有的创作方向。
用自然光影和质感呈现视觉细节
利用自然光线和丰富的表面细节创作逼真场景,呈现木质书桌的纹理、人像的柔和质感或玻璃上的反光。
帮助团队保持视觉风格一致
在多轮编辑和基于参考图制作的不同版本中延续选定的风格。围绕统一的视觉方向制作演示文稿、品牌故事或系列专题内容。
ChatGPT Images 2.6 AI 图像生成器的应用场景
为内容发布、营销和设计创作视觉素材。先考虑受众和展示位置,再根据每张图片需要传达的信息调整画面。
博客和落地页视觉素材制作
围绕主题制作文章封面、网站首屏主视觉和配套插画。调整构图,为标题或页面内容留出空间。
电商产品与营销活动创意素材
从参考照片出发,探索桌面陈列、季节性产品场景和新品发布创意。围绕主推商品安排道具、陈列台面和光线。
社交媒体营销与付费广告
以统一风格制作社媒图片、缩略图和营销活动配图的初稿。先比较不同背景和视觉焦点,再深化喜欢的方向。
概念分镜与创意原型制作
将文字描述的场景转化为分镜图或情绪板概念图。在规划最终设计或拍摄前,先探索角色、场景和取景构图。
如何使用 ChatGPT Images 2.6 AI 图像生成器
从一个创意或一张原始照片开始。生成初稿、检查结果,再完善对项目重要的细节。
输入提示词,明确目标
描述主体、场景、光线和构图。编辑图像时,加入参考照片,并说明想修改的内容和希望保留的细节。
设定方向并生成
从可用选项中选择合适的设置,检查创作说明,然后生成初稿。明确画面主体和预期用途。
完善、重新生成并导出
检查结果,每次修改一处细节。用你喜欢的图像作为参考继续编辑,再下载选定的版本。
ChatGPT Images 社区创作示例
We have a gif for you—ChatGPT Images 2.6 is here. ChatGPT Images 2.6 can turn your prompts and pictures into animated GIFs. You can also keep reference images of yourself, pets, and characters ready to reuse in new creations...no need to upload them again. Go on. Make it
呃,用最新的 GPT Image 2.6 做了个测试。终于,能够直接生成动图了,我直接鹈鹕测试哈哈哈哈😂 谁能想到生图模型现在都可以直接上鹈鹕测试了啊! 说真的在我心里 GPT Image 2.5 应该已经是最强的生图模型了,我甚至不知道他后续还有什么地方可以继续升级了,没想到还是带来了 GPT Image 2.6
GIFS! With ChatGPT Images 2.6, turn a prompt or image into an animated GIF. Adjust the speed, edit individual frames, and share it on repeat.
The GPT Image 2.6 API gives you more control over image complexity and lets you create GIFs. So as fun short experiment - I asked Flare to create a GIF of a man doing a standing backflip. Compare low, medium and high complexity side by side. 👇 Detailed explanation below