{"@context":"https://schema.org","@type":"NewsArticle","generatedAt":"2026-09-28T06:03:00.468Z","headline":"OpenAI 发布 ChatGPT Images 2.5 双图像模型，主打更快生成与局部编辑","description":"OpenAI 发布 ChatGPT Images 2.5，含 GPT-Image-2.5 Flare 和 GPT-Image-2.5 Sunburst 两个模型，主推更自然的光照、更可靠的参考图保留和局部编辑，Flare 延迟较 Images 2.0 降低最多 50%。","url":"https://www.aioga.com/news/cmtu2qhzo0xb2rofp6eaokqdd/","mainEntityOfPage":"https://www.aioga.com/news/cmtu2qhzo0xb2rofp6eaokqdd/","datePublished":"2026-09-09T12:17:58.000Z","dateModified":"2026-09-09T12:17:58.000Z","inLanguage":"zh-CN","publisher":{"@type":"NewsMediaOrganization","name":"Aioga","url":"https://www.aioga.com"},"citation":["https://the-decoder.com/chatgpt-images-2-5-faster-more-precise-but-not-the-same-for-everyone","https://aihot.news/items/cmtu2qhzo0xb2rofp6eaokqdd"],"canonicalUrl":"https://www.aioga.com/news/cmtu2qhzo0xb2rofp6eaokqdd/","directAnswer":{"@type":"Answer","text":"OpenAI 发布 ChatGPT Images 2.5，包含 GPT-Image-2.5 Flare 与 GPT-Image-2.5 Sunburst。来源称其改进光照、纹理、参考图保留和局部编辑，并称 Flare 相比 Images 2.0 延迟最多降低 50%。","url":"https://www.aioga.com/news/cmtu2qhzo0xb2rofp6eaokqdd/","dateCreated":"2026-09-09T12:17:58.000Z","author":{"@type":"Organization","@id":"https://www.aioga.com/authors/aioga-editorial/#editorial-team","name":"Aioga Editorial Team","url":"https://www.aioga.com/authors/aioga-editorial/"}},"evidence":[{"@type":"CreativeWork","name":"the-decoder.com source article","url":"https://the-decoder.com/chatgpt-images-2-5-faster-more-precise-but-not-the-same-for-everyone","datePublished":"2026-09-09T12:17:58.000Z","provider":{"@type":"Organization","name":"the-decoder.com","url":"https://the-decoder.com/chatgpt-images-2-5-faster-more-precise-but-not-the-same-for-everyone"}},{"@type":"CreativeWork","name":"AIHot archive record","url":"https://aihot.news/items/cmtu2qhzo0xb2rofp6eaokqdd","datePublished":"2026-09-09T12:17:58.000Z","provider":{"@type":"Organization","name":"AIHot","url":"https://aihot.news/items/cmtu2qhzo0xb2rofp6eaokqdd"}}],"aggregationSource":"The Decoder：AI News（RSS）","originalPublisher":{"name":"the-decoder.com","url":"https://the-decoder.com/chatgpt-images-2-5-faster-more-precise-but-not-the-same-for-everyone"},"geoDeepAnswer":null,"article":{"id":"cmtu2qhzo0xb2rofp6eaokqdd","slug":"cmtu2qhzo0xb2rofp6eaokqdd","url":"https://www.aioga.com/news/cmtu2qhzo0xb2rofp6eaokqdd/","title":"OpenAI 发布 ChatGPT Images 2.5 双图像模型，主打更快生成与局部编辑","title_en":"","summary":"OpenAI 发布 ChatGPT Images 2.5，含 GPT-Image-2.5 Flare 和 GPT-Image-2.5 Sunburst 两个模型，主推更自然的光照、更可靠的参考图保留和局部编辑，Flare 延迟较 Images 2.0 降低最多 50%。","source":"The Decoder：AI News（RSS）","sourceUrl":"https://the-decoder.com/chatgpt-images-2-5-faster-more-precise-but-not-the-same-for-everyone","aiHotUrl":"https://aihot.news/items/cmtu2qhzo0xb2rofp6eaokqdd","publishedAt":"2026-09-09T12:17:58.000Z","category":"行业动态","score":72,"selected":true,"articleBody":["OpenAI has released two new image models under ChatGPT Images 2.5, promising sharper details, more precise editing, and faster generation. New features like a drawing tool and shareable prompts aim to expand the creative process.","OpenAI says users now generate more than three billion images each week through ChatGPT Images and the GPT-Image models in the API. With ChatGPT Images 2.5, the company is rolling out a reworked generation that promises more natural lighting and finer textures.","The models are meant to preserve subjects from reference photos better and follow editing instructions more reliably across multiple rounds. According to OpenAI, the faster variant cuts image generation latency by up to 50 percent compared to Images 2.0.","For developers, OpenAI brings both models to the API. GPT-Image-2.5 Flare is the default pick for most uses, with higher image quality than GPT-Image-2 at 50 percent lower latency. The stronger model, GPT-Image-2.5 Sunburst, targets more demanding visual work with tighter control over edits, and it needs longer generation times to deliver.","Both models use the same token rates. That's eight dollars per one million image input tokens and 30 dollars per one million output tokens. But since token use varies by model and quality tier, the same rates don't mean the same cost per image. New in Images 2.5 are the \"xhigh\" and \"max\" quality tiers, which go beyond the previous ceiling of \"high.\"","A 1024x1024 image at the \"low\" tier costs about 0.006 dollars, same as the predecessor, while \"high\" runs about 0.053 dollars, and the new \"max\" tier lands at roughly 0.21 dollars with around 7,024 output tokens. That puts the \"max\" tier of Images 2.5 at the same price as the \"high\" tier of GPT-Image-2. Unlike the predecessor, Images 2.5 has no cheaper batch rate so far. In early tests, though, Sunburst usually costs more per image than Flare despite identical token prices, probably because of longer reasoning runs. And unlike last time, OpenAI gives no average price-per-image figure.","It's also unclear how OpenAI routes ChatGPT users between the two models. Neither the announcement nor the documentation says when ChatGPT reaches for the faster Flare or the more precise Sunburst. In the API you can pick the model explicitly. The ChatGPT interface offers no such control yet.","In our tests, the line currently seems to run mainly between Chat and Work. In Work, our prompts really do change only what's asked, no matter the reasoning settings. These targeted edits are the focus of the model improvements. In Chat, though, more details keep shifting in the follow-up images, even with reasoning set to high. Only at the \"6 Pro\" setting does the stronger model sometimes appear to kick in.","As noted, the new model's focus is editing. It's meant to change only the requested elements and leave the rest of an image untouched, even with more complex subjects and backgrounds. In longer conversations, earlier changes should stay consistent without image quality dropping over multiple editing steps. OpenAI shows this with a room redesign, for example. Where the older model changed other details with every edit, the new model stays stable even across several iterations.","A quick test through ChatGPT Work with GPT-6 Astra (Max) shows how well this works. We used the following prompt and then iteratively adjusted one small detail (banana color) and one large one (the big cat).","A hyper-realistic DSLR photo. A monkey holding a pink banana is sitting on a tiger in the foreground. In the background, a HORSE is RIDING AN ASTRONAUT. The astronaut is underneath, like a living “spacesuit horse saddle,” and the HORSE is clearly on top, in control, as the rider. Make it 100% unambiguous: the HORSE is the rider and the ASTRONAUT is being ridden, NOT the other way around. High resolution, sharp focus, realistic lighting.","：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-10_54_33-work-astra-max.png ：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-10_54_33-work-astra-max-yellow.png ：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-10_54_33-work-astra-max-yellow-puma.png Image: GPT-Images-2.5 / Astra (Max)","On a side note, this is probably the best version of a horse riding an astronaut that an OpenAI image model has produced in our tests. Here's how it looked with Image 2.0 (Thinking variant).","Tests in ChatGPT's Chat mode, by contrast, always changed the rest of the image when we adjusted the banana color. Only in \"6 Pro\" mode did the image stay consistent in one run, but not in another. This may change over the course of the week as the rollout continues.","Images 2.5 is also meant to handle complex visual instructions better, deliver more accurate content for real-world information, and work with transparent backgrounds and more demanding layouts. In our last article, for instance, we had ChatGPT turn the piece into an 80s magazine spread. It looked like this.","GPT-Images-2.5 via Astra (Max) also produced a detailed magazine with a sample image based on our monkey-astronaut prompt, in a weaker and a stronger variant, without losing sight of the original instruction for worse quality.","The model corrected an error we planted (GPT-Images-2.5 instead of 2.0 right above the images on the first page) on command, showing how well it preserves the rest of the content, and it did the same for a translation on the simple request \"Create a US English version.\"","：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_25_20-Decoder-Astra-Max-2.5.png ：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_25_20-Decoder-Astra-Max-2.0-fixed.png ：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_25_20-Decoder-Astra-Max-translation.png","Image: GPT-Images-2.5 / Astra (Max)","For comparison, the results through plain ChatGPT Chat are also worth a look, but they show that the weaker model changes other details during translation too. In one case, though, it nailed the current Microsoft CEO better.","：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_35_55-Decoder-Medium.png ：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_36_37-Decoder-Medium-Translate.png ：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-12_42_37-Decoder-Instant.png ：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-12_42_51-Instant-Translation.png","Image: GPT-Images-2.5 / ChatGPT Chat (Medium and Instant)","In ChatGPT, OpenAI is adding several new features alongside the new models. The most important one is called \"Sketch.\" It lets users draw directly in ChatGPT and use the sketch as a visual template for the finished image. You activate it with the \"@Sketch\" command, and OpenAI says it works well for diagrams, room layouts, or posters.","Templates are ready-made prompts meant to make it easier to get started with formats like posters, logos, infographics, thumbnails, illustrations, or ads. They give you a structure instead of a blank canvas and pin down your requirements through targeted follow-up questions. Users can also place comments directly on images and share the prompts they used, so others can try the same idea with their own photos and details. As an example, OpenAI points to a currently viral prompt that generates portraits in 1980s style.","In Arena's text-to-image leaderboard, the new models hold the top two spots for now. GPT-Image-2.5 Sunburst leads with a score of 1421, followed by GPT-Image-2.5 Flare at 1399. Both scores carry a \"Preliminary\" label and rest on still-low vote counts of around 3,100 and 2,900. In third place is the predecessor GPT-Image-2 with 1381 points from about 78,700 votes.","Behind them come Microsoft's mai-image-2.6 (1331), SpaceXAI's grok-imagine-image-2.0 (1315), and several models from Reve, Meta, Google, and Bytedance. Since the ratings for the two new OpenAI models are preliminary, their position could still shift as more votes come in.","For provenance labeling, OpenAI still relies on the C2PA industry standard, which embeds metadata for tracking. To make provenance more resistant, OpenAI also adds an invisible watermark via Google DeepMind's SynthID, in ChatGPT, Codex, and the API. OpenAI says there's no single solution for provenance labeling, which is why it takes a layered approach.","Images 2.5 is available now worldwide for all ChatGPT, ChatGPT Work, and Codex users across desktop, mobile, and web, including the free version. OpenAI says shorter wait times and higher usage limits apply. Our tests suggest the Chat mode currently uses the weaker model.","Stay in the loop on AI. Clear, useful, no fluff.","Follow The Decoder for AI news, background stories and expert analyses.","The Decoder：https://the-decoder.com/"],"articleImages":[{"sourceUrl":"https://the-decoder.com/wp-content/uploads/2026/09/OpenAI-ChatGPT-images-2.5-retrofuturism.webp","alt":"Image description","afterParagraph":0,"url":"/media/articles/cmtu2qhzo0xb2rofp6eaokqdd/1ec872691cc7161f.webp"},{"sourceUrl":"https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-10_54_33-work-astra-max-1200x675.png","alt":"","afterParagraph":10,"url":"/media/articles/cmtu2qhzo0xb2rofp6eaokqdd/754a32255ef7f534.png"},{"sourceUrl":"https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-10_54_33-work-astra-max-yellow-1200x675.png","alt":"","afterParagraph":10,"url":"/media/articles/cmtu2qhzo0xb2rofp6eaokqdd/0d75ded654d0d2d6.png"},{"sourceUrl":"https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-10_54_33-work-astra-max-yellow-puma-1200x675.png","alt":"","afterParagraph":10,"url":"/media/articles/cmtu2qhzo0xb2rofp6eaokqdd/b6134d735883e74e.png"},{"sourceUrl":"https://the-decoder.com/wp-content/uploads/2026/04/astronaut_thinking-1200x900.png","alt":"","afterParagraph":12,"url":"/media/articles/cmtu2qhzo0xb2rofp6eaokqdd/66a57c58d4bf0173.png"},{"sourceUrl":"https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-21.-Apr.-2026-22_11_27-1200x900.png","alt":"","afterParagraph":14,"url":"/media/articles/cmtu2qhzo0xb2rofp6eaokqdd/1abe19e7443f2b39.png"}],"mediaStatus":"ok","articleBodyZh":["OpenAI 已经在 ChatGPT 图像 2.5 下发布了两款新的图像模型，承诺提供更清晰的细节、更精确的编辑以及更快速的生成。像绘图工具和可分享提示这样的新功能旨在扩展创意过程。","OpenAI 表示，用户现在每周通过 ChatGPT Images 和 API 中的 GPT-Image 模型生成超过三十亿张图像。在 ChatGPT 图像 2.5 中，公司推出了重新设计的生成方式，承诺实现更自然的光照和更细腻的纹理。","这些模型旨在更好地保留参考照片中的主题，并在多轮编辑中更可靠地遵循编辑指令。根据 OpenAI 的说法，更快的变体相比 Images 2.0 可将图像生成延迟减少多达 50%。","对于开发者，OpenAI 将这两款模型都带到了 API 中。GPT-Image-2.5 Flare 是大多数用途的默认选项，其图像质量高于 GPT-Image-2，同时延迟降低了 50%。更强大的模型 GPT-Image-2.5 Sunburst 针对要求更高的视觉工作，提供更严格的编辑控制，但需要更长的生成时间以达成效果。","两款模型使用相同的令牌费率。每一百万图像输入令牌 8 美元，每一百万输出令牌 30 美元。但由于模型和质量等级不同，令牌使用量不同，因此相同的费率并不意味着每张图像成本相同。Images 2.5 新增了“xhigh”和“max”质量等级，超越了之前的“high”上限。","一张 1024x1024 的“低”等级图像费用约为 0.006 美元，与前代相同，而“高”等级约为 0.053 美元，新的“max”等级大约为 0.21 美元，输出令牌约为 7,024 个。这使得 Images 2.5 的“max”等级价格与 GPT-Image-2 的“high”等级相同。与前代不同，Images 2.5 目前没有更便宜的批量价格。不过在早期测试中，尽管令牌价格相同，Sunburst 通常每张图像的成本高于 Flare，可能是因为推理运行时间更长。而且与上次不同，OpenAI 没有提供每张图像的平均价格数据。","OpenAI 如何在两个模型之间分配 ChatGPT 用户仍不清楚。公告和文档都没有说明 ChatGPT 什么时候调用更快的 Flare 或更精确的 Sunburst。在 API 中，你可以明确选择模型。ChatGPT 界面尚未提供这种控制。","在我们的测试中，目前的分界线似乎主要在 Chat 和 Work 之间。在 Work 中，我们的提示确实只改变所询问的内容，而不管推理设置如何。这些有针对性的编辑是模型改进的重点。然而在 Chat 中，即使将推理设置为高，后续图像中的细节仍在不断变化。只有在“6 Pro”设置下，更强的模型有时才会开始发挥作用。","如前所述，新模型的重点是编辑。它旨在仅更改请求的元素，并保持图像的其他部分不变，即使面临更复杂的主题和背景。在较长的对话中，早期的更改应保持一致，多次编辑步骤中图像质量不会下降。OpenAI 举了一个房间重新设计的例子显示这一点。旧模型在每次编辑时都会更改其他细节，而新模型即使经历多次迭代也保持稳定。","通过 ChatGPT Work 使用 GPT-6 Astra (Max) 的快速测试显示了这一点的效果。我们使用了以下提示，并逐步调整了一个小细节（香蕉颜色）和一个大细节（大猫）。","一张超现实的 DSLR 照片。一只猴子拿着一根粉色香蕉，坐在前景的老虎上。背景中，一匹马正在骑一名宇航员。宇航员在下面，就像活的“宇航服马鞍”，而马显然在上面，处于掌控地位，作为骑手。确保完全明确：马是骑手，宇航员被骑，而不是相反。高分辨率、焦点清晰、光照逼真。","：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-10_54_33-work-astra-max.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-10_54_33-work-astra-max-yellow.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-10_54_33-work-astra-max-yellow-puma.png 图片：GPT-Images-2.5 / Astra (Max)","顺便说一句，这可能是OpenAI图像模型在我们测试中生成的最佳版本。这是用Image 2.0（思维变体）呈现的样子。","相比之下，ChatGPT聊天模式的测试在调整香蕉颜色时，图像的其余部分都会被改变。只有在“6 Pro”模式下，图像在某次运行中保持一致，而在另一次运行中则不然。随着推广的持续，这一情况可能会在本周内发生变化。","Images 2.5 还旨在更好地处理复杂的视觉指令，提供更准确的现实信息内容，并配合透明背景和更具挑战性的版面设计。例如，在我们上一篇文章中，ChatGPT 把这篇文章变成了 80 年代杂志专题。它看起来是这样的。","通过Astra（Max）的GPT-Images-2.5还制作了一本详细的杂志，包含基于猴子-宇航员提示的示例图像，有较弱和较强的版本，且未因质量下降而偏离原始指令。","模型在指令中纠正了我们设置的一个错误（首页图片上方的 GPT-Images-2.5，而不是 2.0），显示它对其余内容的保存程度，同时对简单请求“创建美式英语版本”的翻译也做了同样的处理。","：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_25_20-Decoder-Astra-Max-2.5.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_25_20-Decoder-Astra-Max-2.0-fixed.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_25_20-Decoder-Astra-Max-translation.png","图片：GPT-Images-2.5 / Astra（Max）","作为对比，普通ChatGPT聊天的结果也值得一看，但它们显示较弱的模型在翻译过程中也会改变其他细节。不过，在一个案例中，它更好地准确地描述了现任Microsoft CEO。","：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_35_55-Decoder-Medium.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_36_37-Decoder-Medium-Translate.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-12_42_37-Decoder-Instant.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-12_42_51-Instant-Translation.png","图片：GPT-Images-2.5 / ChatGPT 聊天（中等和即时）","在 ChatGPT 中，OpenAI 正在为新模型添加几个新功能。其中最重要的功能叫做“草图”。它允许用户在 ChatGPT 中直接绘制，并将草图用作生成最终图像的视觉模板。你可以使用“@Sketch”命令激活它，OpenAI 表示它适用于图表、房间布局或海报等场景。","模板是现成的提示，旨在让用户更容易开始制作海报、标志、信息图、缩略图、插图或广告等格式。它们提供了一个结构，而不是空白画布，并通过针对性的后续问题来确定你的需求。用户还可以直接在图像上发表评论，并分享他们使用的提示，这样其他人也可以用自己的照片和细节尝试相同的创意。作为示例，OpenAI 指出了一个目前流行的提示，它可以生成 1980 年代风格的肖像。","在 Arena 的文本生成图像排行榜中，新模型目前占据前两名。GPT-Image-2.5 Sunburst 以 1421 分领先，其次是 GPT-Image-2.5 Flare，得分为 1399。两个分数都带有“初步”标签，并且依然基于约 3,100 和 2,900 的较低投票数。第三名是前代模型 GPT-Image-2，以约 78,700 票获得 1381 分。","紧随其后的是微软的 mai-image-2.6（1331 分）、SpaceXAI 的 grok-imagine-image-2.0（1315 分）以及 Reve、Meta、谷歌和字节跳动的几个模型。由于这两个新的 OpenAI 模型的评分是初步的，他们的位置随着更多投票的加入仍可能发生变化。","在来源标注方面，OpenAI 仍依赖 C2PA 行业标准，该标准嵌入元数据以进行追踪。为了增强来源的抗性，OpenAI 还通过 Google DeepMind 的 SynthID 在 ChatGPT、Codex 和 API 中添加了隐形水印。OpenAI 表示，没有单一的来源标注方案，这也是他们采用分层方法的原因。","Images 2.5 现已在全球范围内向所有 ChatGPT、ChatGPT Work 和 Codex 用户提供，适用于桌面、移动设备和网页，包括免费版。OpenAI 表示，等待时间更短且使用限额更高。我们的测试表明，聊天模式目前使用的是较弱的模型。","保持对 AI 的关注。清晰、有用、无废话。","关注 The Decoder 获取 AI 新闻、背景故事和专家分析。","解码器：https://the-decoder.com/"],"translationStatus":"translated","bodyOrigin":"source-page","editorial":{"summary":"OpenAI 发布 ChatGPT Images 2.5，包含 GPT-Image-2.5 Flare 与 GPT-Image-2.5 Sunburst。来源称其改进光照、纹理、参考图保留和局部编辑，并称 Flare 相比 Images 2.0 延迟最多降低 50%。","background":"两款模型均已接入 API。Flare 被描述为多数用途的默认选择，Sunburst 面向要求更高的视觉工作。两者采用相同 token 单价，但模型与质量档位会影响单张图片成本；ChatGPT 尚未提供模型选择控制。","viewpoint":"Aioga 判断：此次更新同时覆盖生成速度、参考图保留和编辑指令执行。来源测试显示，Chat 与 Work 中的后续编辑表现并不一致，因此发布说明不足以代表所有使用场景下的实际效果。","implications":"可能影响：API 使用者需要在生成速度、编辑控制和质量档位之间取舍；相同 token 费率不代表每张图片成本相同。ChatGPT 用户目前不足以确认系统何时调用 Flare 或 Sunburst。","nextStep":"后续观察：需要关注 OpenAI 是否说明 ChatGPT 在两款模型之间的路由规则，以及后续材料对不同质量档位的成本、生成延迟和多轮编辑表现如何描述。","evidenceRefs":["title","summary","articleBody","source"],"status":"published","aiGenerated":true,"autoApproved":true,"generatedBy":"aioga-editorial:gpt-5.6-sol","reviewedBy":"aioga-editorial-review:gpt-5.6-sol","generatedAt":"2026-09-09T13:12:59.490Z","sourceHash":"4783da5f72d3f461","review":{"approved":true,"groundedness":97,"clarity":92,"duplicationRisk":18,"blockingIssues":[],"notes":["“ChatGPT 用户目前不足以确认系统何时调用 Flare 或 Sunburst”可更直接地表述为“OpenAI 尚未说明 ChatGPT 何时调用 Flare 或 Sunburst”，以避免将信息披露不足表述为用户能力不足。"]},"validation":{"passed":true,"mode":"ai-auto","revisions":1,"checks":["schema","length","source-attribution","editorial-labels","inference-boundary","low-source-overlap","no-html","independent-ai-review"]}},"tags":["行业动态","The Decoder：AI News（RSS）"],"translations":{"zh-CN":{"title":"OpenAI 发布 ChatGPT Images 2.5 双图像模型，主打更快生成与局部编辑","summary":"OpenAI 发布 ChatGPT Images 2.5，含 GPT-Image-2.5 Flare 和 GPT-Image-2.5 Sunburst 两个模型，主推更自然的光照、更可靠的参考图保留和局部编辑，Flare 延迟较 Images 2.0 降低最多 50%。","category":"行业动态","source":"the-decoder.com","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI 发布 ChatGPT Images 2.5 双图像模型，主打更快生成与局部编辑 - Aioga AI资讯","description":"OpenAI 发布 ChatGPT Images 2.5，含 GPT-Image-2.5 Flare 和 GPT-Image-2.5 Sunburst 两个模型，主推更自然的光照、更可靠的参考图保留和局部编辑，Flare 延迟较 Images 2.0 降低最多 50%。","url":"https://www.aioga.com/news/cmtu2qhzo0xb2rofp6eaokqdd/","articleBody":["OpenAI 已经在 ChatGPT 图像 2.5 下发布了两款新的图像模型，承诺提供更清晰的细节、更精确的编辑以及更快速的生成。像绘图工具和可分享提示这样的新功能旨在扩展创意过程。","OpenAI 表示，用户现在每周通过 ChatGPT Images 和 API 中的 GPT-Image 模型生成超过三十亿张图像。在 ChatGPT 图像 2.5 中，公司推出了重新设计的生成方式，承诺实现更自然的光照和更细腻的纹理。","这些模型旨在更好地保留参考照片中的主题，并在多轮编辑中更可靠地遵循编辑指令。根据 OpenAI 的说法，更快的变体相比 Images 2.0 可将图像生成延迟减少多达 50%。","对于开发者，OpenAI 将这两款模型都带到了 API 中。GPT-Image-2.5 Flare 是大多数用途的默认选项，其图像质量高于 GPT-Image-2，同时延迟降低了 50%。更强大的模型 GPT-Image-2.5 Sunburst 针对要求更高的视觉工作，提供更严格的编辑控制，但需要更长的生成时间以达成效果。","两款模型使用相同的令牌费率。每一百万图像输入令牌 8 美元，每一百万输出令牌 30 美元。但由于模型和质量等级不同，令牌使用量不同，因此相同的费率并不意味着每张图像成本相同。Images 2.5 新增了“xhigh”和“max”质量等级，超越了之前的“high”上限。","一张 1024x1024 的“低”等级图像费用约为 0.006 美元，与前代相同，而“高”等级约为 0.053 美元，新的“max”等级大约为 0.21 美元，输出令牌约为 7,024 个。这使得 Images 2.5 的“max”等级价格与 GPT-Image-2 的“high”等级相同。与前代不同，Images 2.5 目前没有更便宜的批量价格。不过在早期测试中，尽管令牌价格相同，Sunburst 通常每张图像的成本高于 Flare，可能是因为推理运行时间更长。而且与上次不同，OpenAI 没有提供每张图像的平均价格数据。","OpenAI 如何在两个模型之间分配 ChatGPT 用户仍不清楚。公告和文档都没有说明 ChatGPT 什么时候调用更快的 Flare 或更精确的 Sunburst。在 API 中，你可以明确选择模型。ChatGPT 界面尚未提供这种控制。","在我们的测试中，目前的分界线似乎主要在 Chat 和 Work 之间。在 Work 中，我们的提示确实只改变所询问的内容，而不管推理设置如何。这些有针对性的编辑是模型改进的重点。然而在 Chat 中，即使将推理设置为高，后续图像中的细节仍在不断变化。只有在“6 Pro”设置下，更强的模型有时才会开始发挥作用。","如前所述，新模型的重点是编辑。它旨在仅更改请求的元素，并保持图像的其他部分不变，即使面临更复杂的主题和背景。在较长的对话中，早期的更改应保持一致，多次编辑步骤中图像质量不会下降。OpenAI 举了一个房间重新设计的例子显示这一点。旧模型在每次编辑时都会更改其他细节，而新模型即使经历多次迭代也保持稳定。","通过 ChatGPT Work 使用 GPT-6 Astra (Max) 的快速测试显示了这一点的效果。我们使用了以下提示，并逐步调整了一个小细节（香蕉颜色）和一个大细节（大猫）。","一张超现实的 DSLR 照片。一只猴子拿着一根粉色香蕉，坐在前景的老虎上。背景中，一匹马正在骑一名宇航员。宇航员在下面，就像活的“宇航服马鞍”，而马显然在上面，处于掌控地位，作为骑手。确保完全明确：马是骑手，宇航员被骑，而不是相反。高分辨率、焦点清晰、光照逼真。","：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-10_54_33-work-astra-max.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-10_54_33-work-astra-max-yellow.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-10_54_33-work-astra-max-yellow-puma.png 图片：GPT-Images-2.5 / Astra (Max)","顺便说一句，这可能是OpenAI图像模型在我们测试中生成的最佳版本。这是用Image 2.0（思维变体）呈现的样子。","相比之下，ChatGPT聊天模式的测试在调整香蕉颜色时，图像的其余部分都会被改变。只有在“6 Pro”模式下，图像在某次运行中保持一致，而在另一次运行中则不然。随着推广的持续，这一情况可能会在本周内发生变化。","Images 2.5 还旨在更好地处理复杂的视觉指令，提供更准确的现实信息内容，并配合透明背景和更具挑战性的版面设计。例如，在我们上一篇文章中，ChatGPT 把这篇文章变成了 80 年代杂志专题。它看起来是这样的。","通过Astra（Max）的GPT-Images-2.5还制作了一本详细的杂志，包含基于猴子-宇航员提示的示例图像，有较弱和较强的版本，且未因质量下降而偏离原始指令。","模型在指令中纠正了我们设置的一个错误（首页图片上方的 GPT-Images-2.5，而不是 2.0），显示它对其余内容的保存程度，同时对简单请求“创建美式英语版本”的翻译也做了同样的处理。","：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_25_20-Decoder-Astra-Max-2.5.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_25_20-Decoder-Astra-Max-2.0-fixed.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_25_20-Decoder-Astra-Max-translation.png","图片：GPT-Images-2.5 / Astra（Max）","作为对比，普通ChatGPT聊天的结果也值得一看，但它们显示较弱的模型在翻译过程中也会改变其他细节。不过，在一个案例中，它更好地准确地描述了现任Microsoft CEO。","：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_35_55-Decoder-Medium.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-11_36_37-Decoder-Medium-Translate.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-12_42_37-Decoder-Instant.png：https://the-decoder.com/wp-content/uploads/2026/09/ChatGPT-Image-9.-Sept.-2026-12_42_51-Instant-Translation.png","图片：GPT-Images-2.5 / ChatGPT 聊天（中等和即时）","在 ChatGPT 中，OpenAI 正在为新模型添加几个新功能。其中最重要的功能叫做“草图”。它允许用户在 ChatGPT 中直接绘制，并将草图用作生成最终图像的视觉模板。你可以使用“@Sketch”命令激活它，OpenAI 表示它适用于图表、房间布局或海报等场景。","模板是现成的提示，旨在让用户更容易开始制作海报、标志、信息图、缩略图、插图或广告等格式。它们提供了一个结构，而不是空白画布，并通过针对性的后续问题来确定你的需求。用户还可以直接在图像上发表评论，并分享他们使用的提示，这样其他人也可以用自己的照片和细节尝试相同的创意。作为示例，OpenAI 指出了一个目前流行的提示，它可以生成 1980 年代风格的肖像。","在 Arena 的文本生成图像排行榜中，新模型目前占据前两名。GPT-Image-2.5 Sunburst 以 1421 分领先，其次是 GPT-Image-2.5 Flare，得分为 1399。两个分数都带有“初步”标签，并且依然基于约 3,100 和 2,900 的较低投票数。第三名是前代模型 GPT-Image-2，以约 78,700 票获得 1381 分。","紧随其后的是微软的 mai-image-2.6（1331 分）、SpaceXAI 的 grok-imagine-image-2.0（1315 分）以及 Reve、Meta、谷歌和字节跳动的几个模型。由于这两个新的 OpenAI 模型的评分是初步的，他们的位置随着更多投票的加入仍可能发生变化。","在来源标注方面，OpenAI 仍依赖 C2PA 行业标准，该标准嵌入元数据以进行追踪。为了增强来源的抗性，OpenAI 还通过 Google DeepMind 的 SynthID 在 ChatGPT、Codex 和 API 中添加了隐形水印。OpenAI 表示，没有单一的来源标注方案，这也是他们采用分层方法的原因。","Images 2.5 现已在全球范围内向所有 ChatGPT、ChatGPT Work 和 Codex 用户提供，适用于桌面、移动设备和网页，包括免费版。OpenAI 表示，等待时间更短且使用限额更高。我们的测试表明，聊天模式目前使用的是较弱的模型。","保持对 AI 的关注。清晰、有用、无废话。","关注 The Decoder 获取 AI 新闻、背景故事和专家分析。","解码器：https://the-decoder.com/"]},"en":{"title":"OpenAI Releases ChatGPT Images 2.5 Dual-Image Model, Focusing on Faster Generation and Local Editing","summary":"OpenAI released ChatGPT Images 2.5, including two models: GPT-Image-2.5 Flare and GPT-Image-2.5 Sunburst, highlighting more natural lighting, more reliable reference image preservation, and local editing. Flare’s latency is reduced by up to 50% compared with Images 2.0.","category":"Industry","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI Releases ChatGPT Images 2.5 Dual-Image Model, Focusing on Faster Generation and Local Editing - Aioga AI News","description":"OpenAI released ChatGPT Images 2.5, including two models: GPT-Image-2.5 Flare and GPT-Image-2.5 Sunburst, highlighting more natural lighting, more reliable reference image preserva...","url":"https://www.aioga.com/en/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:04:39.817Z"},"ja":{"title":"OpenAIはChatGPT Images 2.5をリリースしました。これは二重画像モデルで、生成速度の向上と部分編集に重点を置いています","summary":"OpenAIはChatGPT Images 2.5をリリースしました。これにはGPT-Image-2.5 FlareとGPT-Image-2.5 Sunburstという2つのモデルが含まれています。より自然な光、より信頼性の高い参照画像保持、部分的な編集に重点を置き、Images 2.0と比べてFlareの遅延は最大50%削減されます。","category":"業界動向","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAIはChatGPT Images 2.5をリリースしました。これは二重画像モデルで、生成速度の向上と部分編集に重点を置いています - Aioga AIニュース","description":"OpenAIはChatGPT Images 2.5をリリースしました。これにはGPT-Image-2.5 FlareとGPT-Image-2.5 Sunburstという2つのモデルが含まれています。より自然な光、より信頼性の高い参照画像保持、部分的な編集に重点を置き、Images 2.0と比べてFlareの遅延は最大50%削減されます。","url":"https://www.aioga.com/ja/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:04:41.712Z"},"ko":{"title":"OpenAI는 더 빠른 생성과 부분 편집에 중점을 둔 이중 이미지 모델인 ChatGPT Images 2.5를 출시했습니다","summary":"OpenAI는 ChatGPT Images 2.5를 출시했으며, 여기에는 GPT-Image-2.5 Flare와 GPT-Image-2.5 Sunburst라는 두 가지 모델이 포함되어 있습니다. 이 모델은 더 자연광, 더 신뢰할 수 있는 참조 이미지 보존, 부분 편집에 중점을 두며, Images 2.0에 비해 Flare 지연 시간이 최대 50%까지 줄어듭니다.","category":"업계 동향","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI는 더 빠른 생성과 부분 편집에 중점을 둔 이중 이미지 모델인 ChatGPT Images 2.5를 출시했습니다 - Aioga AI 뉴스","description":"OpenAI는 ChatGPT Images 2.5를 출시했으며, 여기에는 GPT-Image-2.5 Flare와 GPT-Image-2.5 Sunburst라는 두 가지 모델이 포함되어 있습니다. 이 모델은 더 자연광, 더 신뢰할 수 있는 참조 이미지 보존, 부분 편집에 중점을 두며, Images 2.0에 비해 Flare 지연...","url":"https://www.aioga.com/ko/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:04:50.875Z"},"es":{"title":"OpenAI lanzó ChatGPT Images 2.5, un modelo de doble imagen centrado en una generación más rápida y edición parcial","summary":"OpenAI lanzó ChatGPT Images 2.5, que incluye dos modelos: GPT-Image-2.5 Flare y GPT-Image-2.5 Sunburst. Se centra en una iluminación más natural, una retención de imágenes de referencia más fiable y la edición parcial, con una latencia de Flare reducida hasta en un 50% en comparación con Images 2.0.","category":"Industria","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI lanzó ChatGPT Images 2.5, un modelo de doble imagen centrado en una generación más rápida y edición parcial - Aioga Noticias de IA","description":"OpenAI lanzó ChatGPT Images 2.5, que incluye dos modelos: GPT-Image-2.5 Flare y GPT-Image-2.5 Sunburst. Se centra en una iluminación más natural, una retención de imágenes de refer...","url":"https://www.aioga.com/es/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:04:50.803Z"},"fr":{"title":"OpenAI a publié ChatGPT Images 2.5, un modèle à double image, axé sur une génération plus rapide et un montage partiel","summary":"OpenAI a lancé ChatGPT Images 2.5, qui comprend deux modèles : GPT-Image-2.5 Flare et GPT-Image-2.5 Sunburst. Il met l’accent sur un éclairage plus naturel, une rétention d’image de référence plus fiable et un montage partiel, avec une latence Flare réduite jusqu’à 50 % par rapport à Images 2.0.","category":"Industrie","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI a publié ChatGPT Images 2.5, un modèle à double image, axé sur une génération plus rapide et un montage partiel - Aioga Actualités IA","description":"OpenAI a lancé ChatGPT Images 2.5, qui comprend deux modèles : GPT-Image-2.5 Flare et GPT-Image-2.5 Sunburst. Il met l’accent sur un éclairage plus naturel, une rétention d’image d...","url":"https://www.aioga.com/fr/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:00.320Z"},"de":{"title":"OpenAI veröffentlichte ChatGPT Images 2.5, ein Dual-Image-Modell, das sich auf schnellere Generierung und teilweise Bearbeitung konzentriert","summary":"OpenAI veröffentlichte ChatGPT Images 2.5, das zwei Modelle umfasst: GPT-Image-2.5 Flare und GPT-Image-2.5 Sunburst. Es legt den Fokus auf natürlicheres Licht, zuverlässigere Referenzbildspeicherung und teilweise Bearbeitung, wobei die Flare-Latenz im Vergleich zu Images 2.0 um bis zu 50 % reduziert wurde.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI veröffentlichte ChatGPT Images 2.5, ein Dual-Image-Modell, das sich auf schnellere Generierung und teilweise Bearbeitung konzentriert - Aioga KI-News","description":"OpenAI veröffentlichte ChatGPT Images 2.5, das zwei Modelle umfasst: GPT-Image-2.5 Flare und GPT-Image-2.5 Sunburst. Es legt den Fokus auf natürlicheres Licht, zuverlässigere Refer...","url":"https://www.aioga.com/de/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:00.193Z"},"pt-BR":{"title":"A OpenAI lançou o ChatGPT Images 2.5, um modelo de imagem dupla, focado em geração mais rápida e edição parcial","summary":"A OpenAI lançou o ChatGPT Images 2.5, que inclui dois modelos: GPT-Image-2.5 Flare e GPT-Image-2.5 Sunburst. Ele foca em mais iluminação natural, retenção de imagens de referência mais confiável e edição parcial, com a latência do Flare reduzida em até 50% em comparação com o Images 2.0.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"A OpenAI lançou o ChatGPT Images 2.5, um modelo de imagem dupla, focado em geração mais rápida e edição parcial - Aioga Notícias de IA","description":"A OpenAI lançou o ChatGPT Images 2.5, que inclui dois modelos: GPT-Image-2.5 Flare e GPT-Image-2.5 Sunburst. Ele foca em mais iluminação natural, retenção de imagens de referência...","url":"https://www.aioga.com/pt-BR/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:09.646Z"},"ru":{"title":"OpenAI выпустила ChatGPT Images 2.5 — модель с двумя изображениями, ориентированную на более быструю генерацию и частичное редактирование","summary":"OpenAI выпустила ChatGPT Images 2.5, включающую две модели: GPT-Image-2.5 Flare и GPT-Image-2.5 Sunburst. Она делает акцент на более естественном освещении, более надёжном удержании эталонных изображений и частичном редактировании, при этом задержка Flare снижена до 50% по сравнению с Images 2.0.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI выпустила ChatGPT Images 2.5 — модель с двумя изображениями, ориентированную на более быструю генерацию и частичное редактирование - Aioga Новости ИИ","description":"OpenAI выпустила ChatGPT Images 2.5, включающую две модели: GPT-Image-2.5 Flare и GPT-Image-2.5 Sunburst. Она делает акцент на более естественном освещении, более надёжном удержани...","url":"https://www.aioga.com/ru/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:09.689Z"},"ar":{"title":"أصدرت OpenAI نموذج ChatGPT Images 2.5، وهو نموذج مزدوج الصورة، يركز على التوليد الأسرع والتحرير الجزئي","summary":"أصدرت OpenAI برنامج ChatGPT Images 2.5، الذي يتضمن نموذجين: GPT-Image-2.5 Flare و GPT-Image-2.5 Sunburst. يركز على الإضاءة الطبيعية أكثر، واحتفاظ صور مرجعية أكثر موثوقية، وتحرير جزئي، مع تقليل زمن استجابة التوهج بنسبة تصل إلى 50٪ مقارنة ب Images 2.0.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"أصدرت OpenAI نموذج ChatGPT Images 2.5، وهو نموذج مزدوج الصورة، يركز على التوليد الأسرع والتحرير الجزئي - Aioga أخبار الذكاء الاصطناعي","description":"أصدرت OpenAI برنامج ChatGPT Images 2.5، الذي يتضمن نموذجين: GPT-Image-2.5 Flare و GPT-Image-2.5 Sunburst. يركز على الإضاءة الطبيعية أكثر، واحتفاظ صور مرجعية أكثر موثوقية، وتحرير جز...","url":"https://www.aioga.com/ar/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:18.850Z"},"hi":{"title":"OpenAI ने ChatGPT Images 2.5 जारी किया, जो एक दोहरी छवि वाला मॉडल है, जो तेज़ पीढ़ी और आंशिक संपादन पर ध्यान केंद्रित करता है","summary":"OpenAI ने ChatGPT इमेजेज 2.5 जारी किया, जिसमें दो मॉडल शामिल हैं: GPT-Image-2.5 Flare और GPT-Image-2.5 Sunburst। यह अधिक प्राकृतिक प्रकाश व्यवस्था, अधिक विश्वसनीय संदर्भ छवि प्रतिधारण और आंशिक संपादन पर ध्यान केंद्रित करता है, जिसमें Images 50 की तुलना में Flare विलंबता 2.0% तक कम हो जाती है।","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI ने ChatGPT Images 2.5 जारी किया, जो एक दोहरी छवि वाला मॉडल है, जो तेज़ पीढ़ी और आंशिक संपादन पर ध्यान केंद्रित करता है - Aioga AI समाचार","description":"OpenAI ने ChatGPT इमेजेज 2.5 जारी किया, जिसमें दो मॉडल शामिल हैं: GPT-Image-2.5 Flare और GPT-Image-2.5 Sunburst। यह अधिक प्राकृतिक प्रकाश व्यवस्था, अधिक विश्वसनीय संदर्भ छवि प्रतिध...","url":"https://www.aioga.com/hi/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:19.006Z"},"it":{"title":"OpenAI ha rilasciato ChatGPT Images 2.5, un modello a doppia immagine, focalizzato su una generazione più rapida e un editing parziale","summary":"OpenAI ha rilasciato ChatGPT Images 2.5, che include due modelli: GPT-Image-2.5 Flare e GPT-Image-2.5 Sunburst. Si concentra su una luce più naturale, una ritenzione delle immagini di riferimento più affidabile e un editing parziale, con la latenza Flare ridotta fino al 50% rispetto a Images 2.0.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI ha rilasciato ChatGPT Images 2.5, un modello a doppia immagine, focalizzato su una generazione più rapida e un editing parziale - Aioga Notizie IA","description":"OpenAI ha rilasciato ChatGPT Images 2.5, che include due modelli: GPT-Image-2.5 Flare e GPT-Image-2.5 Sunburst. Si concentra su una luce più naturale, una ritenzione delle immagini...","url":"https://www.aioga.com/it/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:28.228Z"},"nl":{"title":"OpenAI bracht ChatGPT Images 2.5 uit, een dual-image model dat zich richt op snellere generatie en gedeeltelijke bewerking","summary":"OpenAI bracht ChatGPT Images 2.5 uit, dat twee modellen bevat: GPT-Image-2.5 Flare en GPT-Image-2.5 Sunburst. Het richt zich op meer natuurlijk licht, betrouwbaardere referentiebeeldbehoud en gedeeltelijke bewerking, waarbij de Flare-latentie tot 50% is verminderd ten opzichte van Images 2.0.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI bracht ChatGPT Images 2.5 uit, een dual-image model dat zich richt op snellere generatie en gedeeltelijke bewerking - Aioga AI-nieuws","description":"OpenAI bracht ChatGPT Images 2.5 uit, dat twee modellen bevat: GPT-Image-2.5 Flare en GPT-Image-2.5 Sunburst. Het richt zich op meer natuurlijk licht, betrouwbaardere referentiebee...","url":"https://www.aioga.com/nl/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:28.232Z"},"tr":{"title":"OpenAI, daha hızlı üretim ve kısmi düzenlemeye odaklanan çift görüntülü ChatGPT Images 2.5'i piyasaya sürdü","summary":"OpenAI, iki modeli içeren ChatGPT Images 2.5'i yayımladı: GPT-Image-2.5 Flare ve GPT-Image-2.5 Sunburst. Daha doğal ışık, daha güvenilir referans görüntü tutma ve kısmi düzenlemeye odaklanıyor; Flare gecikmesi Images 2.0'a kıyasla %50'ye kadar azalıyor.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI, daha hızlı üretim ve kısmi düzenlemeye odaklanan çift görüntülü ChatGPT Images 2.5'i piyasaya sürdü - Aioga AI Haberleri","description":"OpenAI, iki modeli içeren ChatGPT Images 2.5'i yayımladı: GPT-Image-2.5 Flare ve GPT-Image-2.5 Sunburst. Daha doğal ışık, daha güvenilir referans görüntü tutma ve kısmi düzenlemeye...","url":"https://www.aioga.com/tr/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:37.442Z"},"vi":{"title":"OpenAI đã phát hành ChatGPT Images 2.5, một mô hình hai hình ảnh, tập trung vào việc tạo nhanh hơn và chỉnh sửa một phần","summary":"OpenAI đã phát hành ChatGPT Images 2.5, bao gồm hai mô hình: GPT-Image-2.5 Flare và GPT-Image-2.5 Sunburst. Nó tập trung vào ánh sáng tự nhiên hơn, giữ ảnh tham chiếu đáng tin cậy hơn và chỉnh sửa một phần, với độ trễ Flare giảm tới 50% so với Images 2.0.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI đã phát hành ChatGPT Images 2.5, một mô hình hai hình ảnh, tập trung vào việc tạo nhanh hơn và chỉnh sửa một phần - Tin tức AI Aioga","description":"OpenAI đã phát hành ChatGPT Images 2.5, bao gồm hai mô hình: GPT-Image-2.5 Flare và GPT-Image-2.5 Sunburst. Nó tập trung vào ánh sáng tự nhiên hơn, giữ ảnh tham chiếu đáng tin cậy...","url":"https://www.aioga.com/vi/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:37.612Z"},"id":{"title":"OpenAI merilis ChatGPT Images 2.5, model gambar ganda yang berfokus pada generasi yang lebih cepat dan pengeditan parsial","summary":"OpenAI merilis ChatGPT Images 2.5, yang mencakup dua model: GPT-Image-2.5 Flare dan GPT-Image-2.5 Sunburst. Fokus pada pencahayaan yang lebih alami, retensi referensi gambar yang lebih andal, dan pengeditan parsial, dengan latensi Flare dikurangi hingga 50% dibandingkan Images 2.0.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI merilis ChatGPT Images 2.5, model gambar ganda yang berfokus pada generasi yang lebih cepat dan pengeditan parsial - Berita AI Aioga","description":"OpenAI merilis ChatGPT Images 2.5, yang mencakup dua model: GPT-Image-2.5 Flare dan GPT-Image-2.5 Sunburst. Fokus pada pencahayaan yang lebih alami, retensi referensi gambar yang l...","url":"https://www.aioga.com/id/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:46.616Z"},"th":{"title":"OpenAI เปิดตัว ChatGPT Images 2.5 ซึ่งเป็นโมเดลภาพคู่ที่เน้นการสร้างภาพที่รวดเร็วขึ้นและการแก้ไขบางส่วน","summary":"OpenAI ได้ปล่อย ChatGPT Images 2.5 ซึ่งประกอบด้วยสองรุ่น คือ GPT-Image-2.5 Flare และ GPT-Image-2.5 Sunburst โดยเน้นการใช้แสงธรรมชาติมากขึ้น การเก็บรักษาภาพอ้างอิงที่เชื่อถือได้มากขึ้น และการแก้ไขบางส่วน โดยลดความหน่วงของ Flare ได้สูงสุดถึง 50% เมื่อเทียบกับ Images 2.0","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI เปิดตัว ChatGPT Images 2.5 ซึ่งเป็นโมเดลภาพคู่ที่เน้นการสร้างภาพที่รวดเร็วขึ้นและการแก้ไขบางส่วน - ข่าว AI Aioga","description":"OpenAI ได้ปล่อย ChatGPT Images 2.5 ซึ่งประกอบด้วยสองรุ่น คือ GPT-Image-2.5 Flare และ GPT-Image-2.5 Sunburst โดยเน้นการใช้แสงธรรมชาติมากขึ้น การเก็บรักษาภาพอ้างอิงที่เชื่อถือได้มากข...","url":"https://www.aioga.com/th/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:46.752Z"},"pl":{"title":"OpenAI wypuściło ChatGPT Images 2.5, model z dwoma obrazami, skupiający się na szybszym generowaniu i częściowej edycji","summary":"OpenAI wypuściło ChatGPT Images 2.5, który obejmuje dwa modele: GPT-Image-2.5 Flare oraz GPT-Image-2.5 Sunburst. Skupia się na bardziej naturalnym oświetleniu, bardziej niezawodnym zachowaniu zdjęć referencyjnych oraz częściowej edycji, przy czym opóźnienie Flare jest zmniejszone nawet o 50% w porównaniu do Images 2.0.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"OpenAI wypuściło ChatGPT Images 2.5, model z dwoma obrazami, skupiający się na szybszym generowaniu i częściowej edycji - Aioga Wiadomości AI","description":"OpenAI wypuściło ChatGPT Images 2.5, który obejmuje dwa modele: GPT-Image-2.5 Flare oraz GPT-Image-2.5 Sunburst. Skupia się na bardziej naturalnym oświetleniu, bardziej niezawodnym...","url":"https://www.aioga.com/pl/news/cmtu2qhzo0xb2rofp6eaokqdd/","contentTranslated":true,"sourceHash":"becec6c81e9bd25a","translatedAt":"2026-09-09T13:05:55.851Z"}},"evidenceTier":"verified-news","reviewStatus":"editorial-selected","indexable":true,"editorialCover":""}}