谷歌正式推出并全面开放 Gemini Omni 1.2 Flash 模型 1。该版本在功能上实现显著升级,视频生成长度上限从上一代的 10 秒提升至 15 秒,支持以 1 秒为增量指定 3 到 15 秒之间的任意时长 1。新模型引入多轮对话式编辑功能,允许用户在不重新生成的情况下修改细节;同时增强了参考遵循能力,保持角色、物体和场景在长镜头中的一致性 1。
在成本方面,Gemini Omni 1.2 Flash 的价格较上一代降低 30%,每秒 720p 视频的成本降至 0.07 美元 1。目前,该模型已集成至 Gemini API、Google Cloud Console、Flow、Google Vids 以及 Google Ads 等平台 1。所有由该模型生成的输出内容均带有不可察觉的 SynthID 水印 1。
Google has officially released and made fully available the Gemini Omni 1.2 Flash model, introducing native support for 15-second video generation alongside multi-turn conversational editing and enhanced reference consistency 1. The new model is now integrated into major developer and enterprise platforms, including the Gemini API, Google Cloud Console, Flow, Google Vids, and Google Ads 1.
Pricing for Gemini Omni 1.2 Flash has been reduced by 30% compared to the previous version 1.1, with the cost of generating 720p video dropping to $0.07 per second 1. Technically, the maximum duration for video generation has increased from 10 seconds to 15 seconds, allowing users to specify lengths in one-second increments ranging from 3 to 15 seconds 1. The model also introduces multi-turn conversational editing, which enables users to refine details without regenerating the entire clip 1. Furthermore, reference following capabilities have been strengthened to maintain character, object, and scene consistency across longer shots 1. All output content generated by the model includes an imperceptible SynthID watermark 1.
评论
还没有评论,欢迎留下第一条。