[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-gemini-omni-1-1-flash-production-video":3,"news-related-a169099c-96a0-48d4-9632-09154a4c56de":38},{"id":4,"title":5,"summary":6,"content":7,"original_url":8,"source_id":9,"tags":10,"translations":24,"news_slug":31,"published_at":32,"created_at":33,"modified_at":34,"is_published":35,"publish_type":36,"image_url":14,"view_count":37},"a169099c-96a0-48d4-9632-09154a4c56de","Gemini Omni 1.1 Flash 转向生产级:场景续写 40 秒、首尾帧补间、4K 交付","Google 发布 Gemini Omni 1.1 Flash,面向开发者的生产级视频生成更新:场景续写参考前情从最后 1 秒提升到 10 秒、按 10 秒递增最长 40 秒;新增首尾帧补间、3 秒视频参考、360p 草稿快 60% 便宜约三分之二,可放大至 4K;Adobe、Figma、Runway 同批接入。","8 月 27 日,Google 在官方博客发布了 Gemini Omni 1.1 Flash,并直接给它贴上「production-ready」的标签——一次明确面向开发者的生产级更新。过去一年,视频生成模型的竞争焦点正从「单段片段多惊艳」转向「能不能进真实生产管线」,而这次更新几乎每一项能力都对准这个方向:更长的连续叙事、更精确的镜头控制、更便宜的迭代成本。\n\n## 从最后一秒到 10 秒前情:场景续写撑到 40 秒\n\n场景续写(scene extension)是这次更新的核心:模型从已有视频的结尾无缝继续生成新内容。Omni 1.1 把可参考的前情上下文从此前模型的最后 1 秒提升到 10 秒,视觉一致性与叙事连贯性因此明显改善。操作上按 10 秒一段递增,累计可延长到 40 秒——在短视频语境里,这已经是一段完整情节的长度,而不只是一条素材。\n\n## 首尾帧补间、360p 草稿与 4K 成片\n\n可控性方面,新增首尾帧指定:给出镜头的起始帧和结束帧,模型在两个关键帧之间生成连续视频,官方列出的典型场景是复杂运镜、变焦转场和无缝循环。多模态输入还支持最长 3 秒的视频参考,用于锁定角色与画面的视觉一致性。\n\n工作流设计更显工程味:360p 低清草稿比标准 720p 最多快 60%、成本约三分之一,适合反复试错;定稿后再放大到 1080p 或 4K 交付。这套「草稿-定稿」管线,几乎是把传统 CG 分层渲染的成本思路搬进了生成式视频。\n\n## 接口设计与生态位\n\n开发者可以在 Google AI Studio 直接试用,企业走 Gemini Enterprise Agent Platform 的 Agent Platform API。接口沿用会话续接设计:传上一段视频的 previous_interaction_id,再发一句「Continue the scene」即可续写——等于把「延长叙事」做成了 API 原语,而不是提示词技巧。\n\n生态推进同批到位:Adobe 已把 Omni Flash 集成进 Firefly;Figma Weave 创意总监 Itay Schiff 的评价是,这些能力让团队「从生成视频走向真正执导视频」;Runway 首席创意官确认它嵌入了 Runway 的工作流;GMI Cloud 则点名它在教育与讲解类内容上的可靠性。消费端,Google AI Plus、Pro、Ultra 订阅用户当天起可在 Google Flow 使用,Gemini app 也开放了场景续写。\n\n## 观点:竞争的计量单位变了\n\n比单个功能更值得注意的是,视频模型的竞争正在换计量单位:从「一段惊艳 demo」换到「40 秒连贯叙事 × 三分之一成本草稿 × 4K 交付」。当各家的单段质量逐渐趋同,迭代成本与流程可控性才是开发者选型时真正要算的账。Adobe、Figma、Runway 同批接入也说明,视频生成正在从独立工具变成创作软件里的底层能力。\n\n对做 AI 视频产品的人,问题或许不再是「哪个模型单段效果最好」,而是「哪条管线能让我便宜地试错、稳定地交付」——Omni 1.1 Flash 明显在回答后一个问题(原文:[Google 官方博客](https:\u002F\u002Fblog.google\u002Finnovation-and-ai\u002Ftechnology\u002Fdevelopers-tools\u002Fbuild-with-gemini-omni-1-1-flash\u002F))。","https:\u002F\u002Fblog.google\u002Finnovation-and-ai\u002Ftechnology\u002Fdevelopers-tools\u002Fbuild-with-gemini-omni-1-1-flash\u002F","d884df39-706a-45db-b7ed-371f12e54f1f",[11,15,18,21],{"id":12,"name":13,"slug":13,"description":14,"color":14},"a9524a82-a7c5-4daa-bb4b-a7ee77bb0b94","gemini",null,{"id":16,"name":17,"slug":17,"description":14,"color":14},"8cf7490f-2449-4ba7-be19-61befa0d92b4","google",{"id":19,"name":20,"slug":20,"description":14,"color":14},"7e89b5cc-57db-4f37-bc6d-28919a73931c","model-release",{"id":22,"name":23,"slug":23,"description":14,"color":14},"ebe5dcd1-46b1-4298-b8c2-8e0e2f456e56","video-generation",[25],{"id":26,"lang":27,"title":28,"summary":29,"content":30},"782568c2-47b1-43a9-8a9a-930dc953ac57","en","Gemini Omni 1.1 Flash goes production: 40s scenes, keyframes, 4K","Google shipped Gemini Omni 1.1 Flash, a production-ready update for generative video: scene extension now references 10 seconds of context (up from the final second) and reaches 40 seconds cumulatively; first\u002Flast-frame interpolation, 3-second video references, 60%-faster 360p drafts at a third of the cost, and 4K upscaling also ship. Adobe, Figma and Runway adopted it.","On August 27, Google published Gemini Omni 1.1 Flash on its official blog and labeled it \"production-ready\" — an update explicitly aimed at developers. Over the past year, the battleground for generative video has been shifting from \"how stunning is a single clip\" to \"can this survive a real production pipeline\", and nearly every capability in this release points at that shift: longer continuous storytelling, more precise camera control, and cheaper iteration.\n\n## From the final second to 10 seconds of context: scenes now extend to 40 seconds\n\nScene extension is the centerpiece: the model seamlessly continues generating footage from the end of an existing video. Omni 1.1 raises the prior context the model can reference from just the final second in earlier models to a full 10 seconds, which materially improves visual consistency and narrative adherence. Extensions work in 10-second increments, up to a cumulative total of 40 seconds — in short-video terms, that is the length of an actual plot beat, not a mere stock clip.\n\n## Keyframe interpolation, 360p drafts and 4K finals\n\nOn the control side, you can now specify the first and last frames of a shot; the model generates continuous video between the two keyframes. Google calls out complex camera orbits, zoom transitions and seamless loops as the typical use cases. Multimodal input also accepts up to three seconds of video as a reference, to lock character and scene consistency.\n\nThe workflow engineering is just as telling: 360p draft previews are up to 60% faster than the standard 720p and cost about a third as much, which makes trial-and-error cheap; once a cut is locked, it can be upscaled to 1080p or 4K for delivery. This draft-to-final pipeline essentially ports the layered-cost mentality of traditional CG rendering into generative video.\n\n## API design and where it lands\n\nDevelopers can try it directly in Google AI Studio, and enterprises can build against the Agent Platform API on the Gemini Enterprise Agent Platform. The interface keeps a conversation-continuation design: pass the previous video's `previous_interaction_id` along with a \"Continue the scene\" instruction, and the model extends the footage — narrative extension becomes an API primitive rather than a prompting trick.\n\nEcosystem adoption arrived in the same wave: Adobe has integrated Omni Flash into Firefly; Figma Weave's creative director said these capabilities take teams \"beyond generating videos to truly directing them\"; Runway's chief creative officer confirmed it is embedded in Runway's workflow; and GMI Cloud highlighted its reliability for educational and explanatory content. On the consumer side, the model is available in Google Flow for AI Plus, Pro and Ultra subscribers as of day one, with scene extension also live in the Gemini app.\n\n## Opinion: the unit of competition has changed\n\nMore notable than any single feature is that the competition in video models is changing its unit of account: from \"one impressive demo clip\" to \"40 seconds of coherent narrative × one-third-cost drafts × 4K delivery\". As single-clip quality converges across vendors, iteration cost and pipeline controllability are what developers actually compute when they choose. The same-wave adoption by Adobe, Figma and Runway also signals that video generation is settling in as an underlying capability inside creative software rather than a standalone tool.\n\nFor anyone building AI video products, the question is no longer \"which model renders the best single clip\" but \"which pipeline lets me fail cheaply and deliver reliably\" — and Omni 1.1 Flash is clearly answering the second question (source: [Google's official blog](https:\u002F\u002Fblog.google\u002Finnovation-and-ai\u002Ftechnology\u002Fdevelopers-tools\u002Fbuild-with-gemini-omni-1-1-flash\u002F)).","gemini-omni-1-1-flash-production-video","2026-08-28T23:04:00Z","2026-08-28T23:07:04.784916Z","2026-08-28T23:07:04.784928Z",true,"agent",119,{"items":39},[40,45,50,55,60,65],{"id":41,"title":42,"news_slug":43,"published_at":44},"8303f420-b3f1-485a-aa6f-7775256c84a7","Gemini 3.8 Audio 双发:Live 和 Extended Thinking 把思考+说话压到近实时","gemini-3-8-audio-live-extended-thinking","2026-09-16T03:00:00+00:00",{"id":46,"title":47,"news_slug":48,"published_at":49},"039ff515-68e7-4f11-866a-1da97e26eb45","Gemini 3.8 Live 拿下 S2S 实时语音榜第一","gemini-3-8-live-voice-s2s-number-one","2026-09-15T17:00:00+00:00",{"id":51,"title":52,"news_slug":53,"published_at":54},"1ba7e499-d93d-4566-bbeb-0b762904c0ab","Google 把 Lyria 3.5 装进 Gemini 与公开 API:音乐生成从独立工具变成默认选项","lyria-3-5-gemini-app-api","2026-09-06T23:06:32+00:00",{"id":56,"title":57,"news_slug":58,"published_at":59},"e0a484e4-41e0-4f91-9b9b-a196bbdcf3ba","Gemini 3.8 Flash 双发:同价升级 + Cyber 走可信项目 Fairwind","gemini-3-8-flash-cyber-fairwind-launch","2026-09-03T03:00:00+00:00",{"id":61,"title":62,"news_slug":63,"published_at":64},"9af3dd83-6ed9-498d-9da0-547d917f3e19","语音转文字有了专用模型:Gemini 3.5 Transcribe 上线,出稿快 70%","gemini-35-transcribe-dedicated-asr","2026-08-31T13:30:00+00:00",{"id":66,"title":67,"news_slug":68,"published_at":69},"16856034-439d-4915-aed4-80b42ae09c68","Gemini 3.7 Flash：FrontierCode 43.6%，价格腰斩","gemini-3-7-flash-coding-agent-fast","2026-08-13T09:00:00+00:00"]