跳到主要内容Skip to main content

← 全部作品← All pieces / 模型演示Model demos / 209

模型演示Model demos209 / 210

Gemini Omni 1.2 Flash:15 秒视频与多轮编辑Gemini Omni 1.2 Flash: 15-second video and multi-turn editing

模型:Gemini Omni 1.2 Flash。原始发布页、源码、prompt 与许可都在下方。Model: Gemini Omni 1.2 Flash. Original publisher page, source, prompt and licence are below.

作品在原作者的页面运行。如果画面空白,通常是对方禁止嵌入,点“全屏打开原页”。The piece runs on the original author’s page. If it appears blank, embedding is likely blocked; use “Open original”.

这件作品About this piece

演示什么:官方页展示原生 15 秒视频、更强的参考一致性、多轮对话式编辑、自然语音与更稳的运镜;可以先用 360p/720p 快速草拟,再对细节做一次 follow-up edit,最后升到 1080p/4K,而不是整段重生成。\n官方怎么说:15 秒、720p 下每美元成本比 1.1 低 30%,已在 Gemini API 与 Google Cloud 正式可用;输出带 SynthID 水印。\n边界:官方精选案例,不是同题横评;没有给完整 prompt、参考图与种子;运动、快速剧本变化和复杂物理仍是官方承认的待改进项。What it shows: the official page demonstrates native 15-second video, stronger reference adherence, multi-turn conversational editing, natural speech and steadier camera moves; you can draft at 360p/720p, fix a detail with a follow-up edit, then upscale to 1080p/4K instead of regenerating the whole clip.\nWhat the official page says: 15 seconds at 720p costs 30% less than 1.1, is generally available on the Gemini API and Google Cloud, and outputs carry a SynthID watermark.\nBoundary: curated vendor examples, not a same-prompt head-to-head; full prompts, references and seeds are not provided; fast sports action, rapid script changes and complex physics remain acknowledged gaps.

来源与 promptSource & prompt

官方发布内容;本站只链回官方页面,不复制图片、视频或素材。引用与转载以原厂条款为准。Official release; this site only links to the original page and does not copy images, video or media. Quotation and redistribution follow the vendor’s terms.