{"@context":"https://schema.org","@type":"NewsArticle","generatedAt":"2026-09-01T20:00:49.359Z","headline":"Google DeepMind 为 Gemini 推出 agentic 视频理解功能","description":"Google DeepMind 为 Gemini 3.7 Flash、3.6 Flash 和 3.5 Flash-Lite 推出 agentic video understanding，模型动态扫描视频片段，相比固定帧率处理 token 消耗最多降低 88%，成本最多降低 66%，准确率最多提升 7%。","url":"https://www.aioga.com/news/cmtiy0v6k01dproel63pg10os/","mainEntityOfPage":"https://www.aioga.com/news/cmtiy0v6k01dproel63pg10os/","datePublished":"2026-09-01T17:08:51.000Z","dateModified":"2026-09-01T17:08:51.000Z","inLanguage":"zh-CN","publisher":{"@type":"NewsMediaOrganization","name":"Aioga","url":"https://www.aioga.com"},"citation":["https://deepmind.google/blog/introducing-agentic-video-in-gemini","https://aihot.virxact.com/items/cmtiy0v6k01dproel63pg10os"],"canonicalUrl":"https://www.aioga.com/news/cmtiy0v6k01dproel63pg10os/","directAnswer":{"@type":"Answer","text":"Google DeepMind 宣布为 Gemini 3.7 Flash、3.6 Flash 和 3.5 Flash-Lite 推出 agentic video understanding。该能力可动态搜索、扫描并检查视频片段，结合画面、音频与字幕进行分析。","url":"https://www.aioga.com/news/cmtiy0v6k01dproel63pg10os/","dateCreated":"2026-09-01T17:08:51.000Z","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":"Google DeepMind source article","url":"https://deepmind.google/blog/introducing-agentic-video-in-gemini","datePublished":"2026-09-01T17:08:51.000Z","provider":{"@type":"Organization","name":"Google DeepMind","url":"https://deepmind.google/blog/introducing-agentic-video-in-gemini"}},{"@type":"CreativeWork","name":"AIHot archive record","url":"https://aihot.virxact.com/items/cmtiy0v6k01dproel63pg10os","datePublished":"2026-09-01T17:08:51.000Z","provider":{"@type":"Organization","name":"AIHot","url":"https://aihot.virxact.com/items/cmtiy0v6k01dproel63pg10os"}}],"aggregationSource":"Google DeepMind：Blog（RSS）","originalPublisher":{"name":"Google DeepMind","url":"https://deepmind.google/blog/introducing-agentic-video-in-gemini"},"geoDeepAnswer":null,"article":{"id":"cmtiy0v6k01dproel63pg10os","slug":"cmtiy0v6k01dproel63pg10os","url":"https://www.aioga.com/news/cmtiy0v6k01dproel63pg10os/","title":"Google DeepMind 为 Gemini 推出 agentic 视频理解功能","title_en":"","summary":"Google DeepMind 为 Gemini 3.7 Flash、3.6 Flash 和 3.5 Flash-Lite 推出 agentic video understanding，模型动态扫描视频片段，相比固定帧率处理 token 消耗最多降低 88%，成本最多降低 66%，准确率最多提升 7%。","source":"Google DeepMind：Blog（RSS）","sourceUrl":"https://deepmind.google/blog/introducing-agentic-video-in-gemini","aiHotUrl":"https://aihot.virxact.com/items/cmtiy0v6k01dproel63pg10os","publishedAt":"2026-09-01T17:08:51.000Z","category":"行业动态","score":72,"selected":true,"articleBody":["Today, we’re launching agentic video understanding：https://ai.google.dev/gemini-api/docs/video-understanding#agentic-video-understanding across our latest models: Gemini 3.7 Flash, 3.6 Flash and 3.5 Flash-Lite. This new capability improves accuracy while dramatically reducing token usage and costs for video analysis. Similar to agentic vision：https://blog.google/innovation-and-ai/technology/developers-tools/agentic-vision-gemini-3-flash/, which combines code execution with Gemini models’ native image understanding, agentic video understanding uses Gemini’s native video tools to improve performance and unlock new capabilities for video processing like sub-second moment retrieval, more accurate anomaly detection, precise counting and more.","The feature is available today for video uploads and YouTube videos via the Gemini API in Google AI Studio and the Gemini Enterprise Agent Platform.","Unlike current ‘static’ processing, where the model ingests the video at a fixed frames-per-second rate (default 1 FPS, adjustable via API), agentic video understanding pairs the model’s core reasoning with native video tools to dynamically search, scan, and inspect target video segments across visual frames, audio, and transcripts. Across standard video analysis benchmarks, Gemini models with agentic video understanding reduce analysis costs by up to 66% and token consumption by up to 88%, while improving accuracy by up to 7%.","These efficiency gains are especially pronounced on long-form video (from 10-minute how-to guides to 90-minute lectures and multi-hour recordings), where static processing forces developers to choose between high token costs or techniques that drop critical details.","Activating agentic video understanding drops token consumption by up to 88% and boosts accuracy by up to 7% with Gemini 3.7 Flash.","While these gains span all three supported models, Gemini 3.7 Flash with agentic understanding offers the best possible quality overall and the best combination of quality and cost efficiency, putting it at the accuracy-to-cost pareto frontier among tested models for video understanding.","Using agentic video understanding places Gemini 3.7 Flash at the accuracy-to-cost pareto frontier for video analysis.","Instead of static processing where the model ingests media streams at a fixed frame rate, agentic video understanding enables Gemini to take an active, goal-directed role in determining what to watch, at what speed, and through which modality (frames, audio, or transcript), fetching only the moments and signals needed. While developers could previously do this manually, with agentic video understanding, Gemini can accomplish it through an agentic loop, invoking an internal tool to load the relevant part of the video file, significantly reducing development overheads.","Agentic video understanding transforms how developers can process long-form video content across a variety of demanding applications.","Token-efficient long-form video analysis","See how Gemini 3.7 Flash performs with and without agentic video understanding on LongVideoBench, a long-form video understanding benchmark. Notice the large token reductions and accuracy improvements.","Accurate fast action analysis with dynamic FPS","With agentic video understanding, 3.7 Flash is able to accurately count a fast-paced movement by scanning and rewatching the video at different frames per second, as needed.","Token-efficient needle-in-a-haystack search","Using agentic video understanding, Gemini 3.7 is able to accurately answer complex questions based on the content of the video while consuming a significantly lower number of tokens compared to static analysis.","Many of our early access partners saw strong performance while testing with agentic video understanding. Here’s what they have to say:","Agentic video understanding is available via the Gemini API in Google AI Studio：https://ai.google.dev/gemini-api/docs/video-understanding#agentic-video-understanding and Gemini Enterprise Agent Platform：https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/capabilities/video-understanding, launching across Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite. It uses standard Gemini API token pricing with no additional feature fee.","To enable it, simply set processing to \"agentic\" in the API configuration. Read our developer guide：http://ai.dev/learn/agentic-video-understanding-with-gemini to get more insights into the feature and how to get started.","We are also bringing the efficiency and quality improvements of agentic video understanding to billions of users across Google products. The feature will roll out to all users in the Gemini app across Flash and Flash-Lite models soon. And in the coming months, agentic video understanding will also power YouTube's ‘Ask YouTube’：https://support.google.com/youtube/answer/14110396?hl=en&co=GENIE.Platform%3DAndroid feature on the video watch page, leveraging Gemini to deliver higher-quality answers grounded in the visuals.","Acknowledgement for their contribution to this work: Sergi Caelles, Filip Pavetić, Ahmet Iscen, Suhas Yogin, and the Agentic Vision team.","Check your inbox to confirm your subscription."],"articleImages":[],"mediaStatus":"none","articleBodyZh":["今天，我们推出了自主视频理解（Agentic Video Understanding）：https://ai.google.dev/gemini-api/docs/video-understanding#agentic-video-understanding，适用于我们最新的模型：Gemini 3.7 Flash、3.6 Flash 和 3.5 Flash-Lite。这一新功能在显著降低视频分析的 token 使用量和成本的同时，提升了准确性。类似于自主视觉（Agentic Vision）：https://blog.google/innovation-and-ai/technology/developers-tools/agentic-vision-gemini-3-flash/，该技术将代码执行与 Gemini 模型的原生图像理解结合，自主视频理解使用 Gemini 的原生视频工具来提升性能，并为视频处理解锁新功能，例如亚秒级片段检索、更准确的异常检测、精确计数等。","该功能现已通过 Google AI Studio 和 Gemini 企业代理平台的 Gemini API 对视频上传和 YouTube 视频开放。","与当前的“静态”处理不同，静态处理模式下模型以固定帧率（默认 1 FPS，可通过 API 调整）摄取视频，自主视频理解将模型的核心推理与原生视频工具相结合，动态搜索、扫描和检查目标视频片段的视觉帧、音频和转录文本。在标准视频分析基准测试中，配备自主视频理解的 Gemini 模型可将分析成本降低高达 66%，token 消耗降低高达 88%，同时准确率提升高达 7%。","这些效率提升在长视频上尤其明显（从 10 分钟的操作指南到 90 分钟的讲座以及数小时的录制），因为静态处理迫使开发者在高 token 成本和可能丢失关键细节的处理技术之间做出选择。","启用自主视频理解后，使用 Gemini 3.7 Flash 的 token 消耗可降低高达 88%，准确率提升可达 7%。","虽然这些提升适用于所有三种支持的模型，但配备自主理解的 Gemini 3.7 Flash 提供了整体最佳质量及最佳质量与成本效率的组合，使其在视频理解的测试模型中处于准确率与成本的帕累托前沿。","使用自主视频理解使 Gemini 3.7 Flash 在视频分析中达到了准确率与成本的帕累托前沿。","与模型以固定帧率静态处理媒体流不同，智能视频理解使 Gemini 能够主动、目标导向地决定观看内容的方式、速度以及通过哪种模态（帧、音频或文字记录）观看，只提取所需的时刻和信号。虽然开发者以前可以手动完成此操作，但通过智能视频理解，Gemini 可以通过智能循环调用内部工具加载视频文件的相关部分，从而显著减少开发开销。","智能视频理解改变了开发者在各种高要求应用中处理长视频内容的方式。","高效令牌长视频分析","看看 Gemini 3.7 Flash 在 LongVideoBench（长视频理解基准）上有无智能视频理解的表现。注意令牌的大幅减少以及准确性的提升。","动态帧速下的快速动作精准分析","通过智能视频理解，3.7 Flash 可以通过按需扫描和重放不同帧率的视频，准确统计快速动作的次数。","高效令牌中的大海捞针搜索","使用智能视频理解，Gemini 3.7 能够基于视频内容准确回答复杂问题，同时相比静态分析消耗的令牌显著减少。","我们的许多早期访问合作伙伴在测试智能视频理解时看到了强劲的性能表现。他们这样说：","智能视频理解可通过 Google AI Studio 的 Gemini API 间接访问：https://ai.google.dev/gemini-api/docs/video-understanding#agentic-video-understanding 以及 Gemini 企业代理平台：https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/capabilities/video-understanding，已在 Gemini 3.7 Flash、3.6 Flash 和 3.5 Flash-Lite 上推出。它使用标准的 Gemini API 令牌计费，无额外功能费用。","启用方法简单，只需在 API 配置中将处理模式设置为“agentic”。阅读我们的开发者指南：http://ai.dev/learn/agentic-video-understanding-with-gemini 以获取功能更多见解及入门方法。","我们还将代理视频理解的效率和质量提升带给谷歌产品中的数十亿用户。该功能将很快在 Gemini 应用的 Flash 和 Flash-Lite 模型中向所有用户推出。在接下来的几个月中，代理视频理解还将为 YouTube 的“Ask YouTube”: https://support.google.com/youtube/answer/14110396?hl=en&co=GENIE.Platform%3DAndroid 功能提供支持，该功能位于视频观看页面，利用 Gemini 提供基于视觉内容的更高质量答案。","感谢以下人员对这项工作的贡献：Sergi Caelles、Filip Pavetić、Ahmet Iscen、Suhas Yogin 以及代理视觉团队。","请检查您的收件箱以确认您的订阅。"],"translationStatus":"translated","bodyOrigin":"source-page","editorial":{"summary":"Google DeepMind 宣布为 Gemini 3.7 Flash、3.6 Flash 和 3.5 Flash-Lite 推出 agentic video understanding。该能力可动态搜索、扫描并检查视频片段，结合画面、音频与字幕进行分析。","background":"该功能面向 Gemini API 中的视频上传和 YouTube 视频，并可在 Google AI Studio 与 Gemini Enterprise Agent Platform 使用。相较固定帧率处理，官方称其最高可减少 88% 的 token 消耗、降低 66% 的分析成本，并提升 7% 的准确率。","viewpoint":"Aioga 判断：这项更新的核心变化是把视频分析从固定采样转向目标导向的动态检索。来源显示，模型可通过内部工具选择观看时段、速度和信息模态，但实际效果仍取决于具体视频与任务。","implications":"可能影响：长视频分析的成本与细节取舍可能得到改善，但最高幅度指标不足以代表所有场景的实际结果。开发者需要评估模型版本、视频类型和任务目标，不能仅依据基准数据推断上线收益。","nextStep":"后续观察：应关注 Gemini API 文档对调用方式、支持范围和计费口径的进一步说明，并观察不同长度、模态组合及异常检测、精确计数等任务中的实际表现。","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":"2026-09-01T18:11:25.677Z","sourceHash":"5db88eb60542883d","review":{"approved":true,"groundedness":94,"clarity":93,"duplicationRisk":10,"blockingIssues":[],"notes":["“字幕”建议改为“文字记录”或“转录文本”，以更准确对应来源中的 transcripts；来源并未限定为字幕。","“实际效果仍取决于具体视频与任务”属于合理的限定性判断，建议保留为观点或表述为“实际效果可能因具体视频与任务而异”。"]},"validation":{"passed":true,"mode":"ai-auto","revisions":0,"checks":["schema","length","source-attribution","editorial-labels","inference-boundary","low-source-overlap","no-html","independent-ai-review"]}},"tags":["行业动态","Google DeepMind：Blog（RSS）"],"translations":{"zh-CN":{"title":"Google DeepMind 为 Gemini 推出 agentic 视频理解功能","summary":"Google DeepMind 为 Gemini 3.7 Flash、3.6 Flash 和 3.5 Flash-Lite 推出 agentic video understanding，模型动态扫描视频片段，相比固定帧率处理 token 消耗最多降低 88%，成本最多降低 66%，准确率最多提升 7%。","category":"行业动态","source":"Google DeepMind","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind 为 Gemini 推出 agentic 视频理解功能 - Aioga AI资讯","description":"Google DeepMind 为 Gemini 3.7 Flash、3.6 Flash 和 3.5 Flash-Lite 推出 agentic video understanding，模型动态扫描视频片段，相比固定帧率处理 token 消耗最多降低 88%，成本最多降低 66%，准确率最多提升 7%。","url":"https://www.aioga.com/news/cmtiy0v6k01dproel63pg10os/","articleBody":["今天，我们推出了自主视频理解（Agentic Video Understanding）：https://ai.google.dev/gemini-api/docs/video-understanding#agentic-video-understanding，适用于我们最新的模型：Gemini 3.7 Flash、3.6 Flash 和 3.5 Flash-Lite。这一新功能在显著降低视频分析的 token 使用量和成本的同时，提升了准确性。类似于自主视觉（Agentic Vision）：https://blog.google/innovation-and-ai/technology/developers-tools/agentic-vision-gemini-3-flash/，该技术将代码执行与 Gemini 模型的原生图像理解结合，自主视频理解使用 Gemini 的原生视频工具来提升性能，并为视频处理解锁新功能，例如亚秒级片段检索、更准确的异常检测、精确计数等。","该功能现已通过 Google AI Studio 和 Gemini 企业代理平台的 Gemini API 对视频上传和 YouTube 视频开放。","与当前的“静态”处理不同，静态处理模式下模型以固定帧率（默认 1 FPS，可通过 API 调整）摄取视频，自主视频理解将模型的核心推理与原生视频工具相结合，动态搜索、扫描和检查目标视频片段的视觉帧、音频和转录文本。在标准视频分析基准测试中，配备自主视频理解的 Gemini 模型可将分析成本降低高达 66%，token 消耗降低高达 88%，同时准确率提升高达 7%。","这些效率提升在长视频上尤其明显（从 10 分钟的操作指南到 90 分钟的讲座以及数小时的录制），因为静态处理迫使开发者在高 token 成本和可能丢失关键细节的处理技术之间做出选择。","启用自主视频理解后，使用 Gemini 3.7 Flash 的 token 消耗可降低高达 88%，准确率提升可达 7%。","虽然这些提升适用于所有三种支持的模型，但配备自主理解的 Gemini 3.7 Flash 提供了整体最佳质量及最佳质量与成本效率的组合，使其在视频理解的测试模型中处于准确率与成本的帕累托前沿。","使用自主视频理解使 Gemini 3.7 Flash 在视频分析中达到了准确率与成本的帕累托前沿。","与模型以固定帧率静态处理媒体流不同，智能视频理解使 Gemini 能够主动、目标导向地决定观看内容的方式、速度以及通过哪种模态（帧、音频或文字记录）观看，只提取所需的时刻和信号。虽然开发者以前可以手动完成此操作，但通过智能视频理解，Gemini 可以通过智能循环调用内部工具加载视频文件的相关部分，从而显著减少开发开销。","智能视频理解改变了开发者在各种高要求应用中处理长视频内容的方式。","高效令牌长视频分析","看看 Gemini 3.7 Flash 在 LongVideoBench（长视频理解基准）上有无智能视频理解的表现。注意令牌的大幅减少以及准确性的提升。","动态帧速下的快速动作精准分析","通过智能视频理解，3.7 Flash 可以通过按需扫描和重放不同帧率的视频，准确统计快速动作的次数。","高效令牌中的大海捞针搜索","使用智能视频理解，Gemini 3.7 能够基于视频内容准确回答复杂问题，同时相比静态分析消耗的令牌显著减少。","我们的许多早期访问合作伙伴在测试智能视频理解时看到了强劲的性能表现。他们这样说：","智能视频理解可通过 Google AI Studio 的 Gemini API 间接访问：https://ai.google.dev/gemini-api/docs/video-understanding#agentic-video-understanding 以及 Gemini 企业代理平台：https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/capabilities/video-understanding，已在 Gemini 3.7 Flash、3.6 Flash 和 3.5 Flash-Lite 上推出。它使用标准的 Gemini API 令牌计费，无额外功能费用。","启用方法简单，只需在 API 配置中将处理模式设置为“agentic”。阅读我们的开发者指南：http://ai.dev/learn/agentic-video-understanding-with-gemini 以获取功能更多见解及入门方法。","我们还将代理视频理解的效率和质量提升带给谷歌产品中的数十亿用户。该功能将很快在 Gemini 应用的 Flash 和 Flash-Lite 模型中向所有用户推出。在接下来的几个月中，代理视频理解还将为 YouTube 的“Ask YouTube”: https://support.google.com/youtube/answer/14110396?hl=en&co=GENIE.Platform%3DAndroid 功能提供支持，该功能位于视频观看页面，利用 Gemini 提供基于视觉内容的更高质量答案。","感谢以下人员对这项工作的贡献：Sergi Caelles、Filip Pavetić、Ahmet Iscen、Suhas Yogin 以及代理视觉团队。","请检查您的收件箱以确认您的订阅。"]},"en":{"title":"Google DeepMind launches agentic video understanding feature for Gemini","summary":"Google DeepMind has launched agentic video understanding for Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite. The model dynamically scans video clips, reducing token consumption by up to 88% compared to fixed frame rate processing, cutting costs by up to 66%, and increasing accuracy by up to 7%.","category":"Industry","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind launches agentic video understanding feature for Gemini - Aioga AI News","description":"Google DeepMind has launched agentic video understanding for Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite. The model dynamically scans video clips, reducing token consumption by...","url":"https://www.aioga.com/en/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:02:32.617Z"},"ja":{"title":"Google DeepMindはGemini向けにエージェント型ビデオ理解機能を導入","summary":"Google DeepMind は Gemini 3.7 Flash、3.6 Flash、3.5 Flash-Lite に対してエージェント型ビデオ理解を導入しました。モデルは動画クリップを動的にスキャンし、固定フレームレートで処理する場合と比べて、トークン消費を最大で88%削減し、コストを最大で66%削減、精度を最大で7%向上させます。","category":"業界動向","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMindはGemini向けにエージェント型ビデオ理解機能を導入 - Aioga AIニュース","description":"Google DeepMind は Gemini 3.7 Flash、3.6 Flash、3.5 Flash-Lite に対してエージェント型ビデオ理解を導入しました。モデルは動画クリップを動的にスキャンし、固定フレームレートで処理する場合と比べて、トークン消費を最大で88%削減し、コストを最大で66%削減、精度を最大で7%向上させます。","url":"https://www.aioga.com/ja/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:02:44.223Z"},"ko":{"title":"Google DeepMind가 Gemini를 위해 에이전트형 동영상 이해 기능을 출시했습니다","summary":"Google DeepMind는 Gemini 3.7 Flash, 3.6 Flash 및 3.5 Flash-Lite를 위해 agentic video understanding을 출시했으며, 모델이 비디오 클립을 동적으로 스캔하여 고정 프레임 속도 대비 토큰 소모를 최대 88%까지 줄이고, 비용을 최대 66%까지 줄이며, 정확도를 최대 7%까지 향상시켰습니다.","category":"업계 동향","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind가 Gemini를 위해 에이전트형 동영상 이해 기능을 출시했습니다 - Aioga AI 뉴스","description":"Google DeepMind는 Gemini 3.7 Flash, 3.6 Flash 및 3.5 Flash-Lite를 위해 agentic video understanding을 출시했으며, 모델이 비디오 클립을 동적으로 스캔하여 고정 프레임 속도 대비 토큰 소모를 최대 88%까지 줄이고, 비용을 최대 66%까지 줄이며, 정확도를...","url":"https://www.aioga.com/ko/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:03:24.262Z"},"es":{"title":"Google DeepMind lanza la función de comprensión de video agente para Gemini","summary":"Google DeepMind ha lanzado la comprensión de video agentic para Gemini 3.7 Flash, 3.6 Flash y 3.5 Flash-Lite. El modelo escanea dinámicamente los fragmentos de video, reduciendo hasta un 88% el consumo de tokens en comparación con el procesamiento a tasa de cuadros fija, disminuyendo hasta un 66% los costos y aumentando la precisión hasta en un 7%.","category":"Industria","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind lanza la función de comprensión de video agente para Gemini - Aioga Noticias de IA","description":"Google DeepMind ha lanzado la comprensión de video agentic para Gemini 3.7 Flash, 3.6 Flash y 3.5 Flash-Lite. El modelo escanea dinámicamente los fragmentos de video, reduciendo ha...","url":"https://www.aioga.com/es/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:03:20.947Z"},"fr":{"title":"Google DeepMind lance la fonctionnalité de compréhension vidéo agentic pour Gemini","summary":"Google DeepMind a lancé la compréhension vidéo agentive pour Gemini 3.7 Flash, 3.6 Flash et 3.5 Flash-Lite. Le modèle analyse dynamiquement les séquences vidéo, réduisant la consommation de tokens jusqu'à 88 % par rapport au traitement à fréquence d'image fixe, diminuant les coûts jusqu'à 66 % et augmentant la précision jusqu'à 7 %.","category":"Industrie","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind lance la fonctionnalité de compréhension vidéo agentic pour Gemini - Aioga Actualités IA","description":"Google DeepMind a lancé la compréhension vidéo agentive pour Gemini 3.7 Flash, 3.6 Flash et 3.5 Flash-Lite. Le modèle analyse dynamiquement les séquences vidéo, réduisant la consom...","url":"https://www.aioga.com/fr/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:04:02.268Z"},"de":{"title":"Google DeepMind führt agentenbasierte Videoverstehensfunktion für Gemini ein","summary":"Google DeepMind hat für Gemini 3.7 Flash, 3.6 Flash und 3.5 Flash-Lite agentisches Videoverständnis eingeführt. Das Modell scannt Videosegmente dynamisch, wodurch im Vergleich zur Verarbeitung mit fester Bildrate der Token-Verbrauch um bis zu 88 % reduziert, die Kosten um bis zu 66 % gesenkt und die Genauigkeit um bis zu 7 % erhöht werden.","category":"行业动态","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind führt agentenbasierte Videoverstehensfunktion für Gemini ein - Aioga KI-News","description":"Google DeepMind hat für Gemini 3.7 Flash, 3.6 Flash und 3.5 Flash-Lite agentisches Videoverständnis eingeführt. Das Modell scannt Videosegmente dynamisch, wodurch im Vergleich zur...","url":"https://www.aioga.com/de/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:04:05.054Z"},"pt-BR":{"title":"Google DeepMind lança recurso de compreensão de vídeo agentic para Gemini","summary":"O Google DeepMind lançou o entendimento de vídeo agentic para o Gemini 3.7 Flash, 3.6 Flash e 3.5 Flash-Lite, onde o modelo realiza varredura dinâmica nos clipes de vídeo, reduzindo o consumo de tokens em até 88% em comparação com o processamento de taxa de quadros fixa, diminuindo os custos em até 66% e aumentando a precisão em até 7%.","category":"行业动态","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind lança recurso de compreensão de vídeo agentic para Gemini - Aioga Notícias de IA","description":"O Google DeepMind lançou o entendimento de vídeo agentic para o Gemini 3.7 Flash, 3.6 Flash e 3.5 Flash-Lite, onde o modelo realiza varredura dinâmica nos clipes de vídeo, reduzind...","url":"https://www.aioga.com/pt-BR/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:04:39.892Z"},"ru":{"title":"Google DeepMind представил функцию агентного видеопонимания для Gemini","summary":"Google DeepMind представила агентное понимание видео для Gemini 3.7 Flash, 3.6 Flash и 3.5 Flash-Lite, модель динамически сканирует видеоклипы, по сравнению с обработкой токенов с фиксированной частотой кадров, расход токенов сокращается максимум на 88%, затраты снижаются максимум на 66%, а точность увеличивается максимум на 7%.","category":"行业动态","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind представил функцию агентного видеопонимания для Gemini - Aioga Новости ИИ","description":"Google DeepMind представила агентное понимание видео для Gemini 3.7 Flash, 3.6 Flash и 3.5 Flash-Lite, модель динамически сканирует видеоклипы, по сравнению с обработкой токенов с...","url":"https://www.aioga.com/ru/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:04:47.336Z"},"ar":{"title":"طرحت Google DeepMind ميزة الفهم الفيديوي الوكالي لـ Gemini","summary":"قدمت Google DeepMind فهم الفيديو الوكلي لـ Gemini 3.7 Flash و 3.6 Flash و 3.5 Flash-Lite، حيث يقوم النموذج بمسح مقاطع الفيديو ديناميكيًا، وبالمقارنة مع معالجة الرموز بمعدل إطار ثابت، يتم تقليل الاستهلاك بنسبة تصل إلى 88%، وتقليل التكلفة بنسبة تصل إلى 66%، وزيادة الدقة بنسبة تصل إلى 7%.","category":"行业动态","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"طرحت Google DeepMind ميزة الفهم الفيديوي الوكالي لـ Gemini - Aioga أخبار الذكاء الاصطناعي","description":"قدمت Google DeepMind فهم الفيديو الوكلي لـ Gemini 3.7 Flash و 3.6 Flash و 3.5 Flash-Lite، حيث يقوم النموذج بمسح مقاطع الفيديو ديناميكيًا، وبالمقارنة مع معالجة الرموز بمعدل إطار ثاب...","url":"https://www.aioga.com/ar/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:05:30.483Z"},"hi":{"title":"Google DeepMind ने Gemini के लिए एजेंटिक वीडियो समझने की सुविधा लॉन्च की","summary":"Google DeepMind ने Gemini 3.7 Flash, 3.6 Flash और 3.5 Flash-Lite के लिए एजेंटिक वीडियो समझ पेश किया, मॉडल वीडियो क्लिप्स को डायनेमिक रूप से स्कैन करता है, और स्थिर फ्रेम दर के मुकाबले टोकन की खपत अधिकतम 88% तक कम करता है, लागत अधिकतम 66% तक कम करता है, और सटीकता अधिकतम 7% तक बढ़ा देता है।","category":"行业动态","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind ने Gemini के लिए एजेंटिक वीडियो समझने की सुविधा लॉन्च की - Aioga AI समाचार","description":"Google DeepMind ने Gemini 3.7 Flash, 3.6 Flash और 3.5 Flash-Lite के लिए एजेंटिक वीडियो समझ पेश किया, मॉडल वीडियो क्लिप्स को डायनेमिक रूप से स्कैन करता है, और स्थिर फ्रेम दर के मुका...","url":"https://www.aioga.com/hi/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:05:32.044Z"},"it":{"title":"Google DeepMind lancia la funzione di comprensione video agentic per Gemini","summary":"Google DeepMind ha lanciato la comprensione video agentica per Gemini 3.7 Flash, 3.6 Flash e 3.5 Flash-Lite, con il modello che esegue la scansione dinamica dei frammenti video; rispetto all'elaborazione a frame rate fisso, il consumo di token può diminuire fino all'88%, i costi fino al 66% e la precisione può aumentare fino al 7%.","category":"行业动态","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind lancia la funzione di comprensione video agentic per Gemini - Aioga Notizie IA","description":"Google DeepMind ha lanciato la comprensione video agentica per Gemini 3.7 Flash, 3.6 Flash e 3.5 Flash-Lite, con il modello che esegue la scansione dinamica dei frammenti video; ri...","url":"https://www.aioga.com/it/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:06:14.119Z"},"nl":{"title":"Google DeepMind introduceert agentische videobegripsfunctionaliteit voor Gemini","summary":"Google DeepMind heeft agentische videoverwerking geïntroduceerd voor Gemini 3.7 Flash, 3.6 Flash en 3.5 Flash-Lite. Het model scant videofragmenten dynamisch en vergeleken met verwerking met een vaste framesnelheid kan het tokenverbruik met maximaal 88% worden verminderd, de kosten met maximaal 66% worden verlaagd en de nauwkeurigheid met maximaal 7% worden verhoogd.","category":"行业动态","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind introduceert agentische videobegripsfunctionaliteit voor Gemini - Aioga AI-nieuws","description":"Google DeepMind heeft agentische videoverwerking geïntroduceerd voor Gemini 3.7 Flash, 3.6 Flash en 3.5 Flash-Lite. Het model scant videofragmenten dynamisch en vergeleken met verw...","url":"https://www.aioga.com/nl/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:06:08.949Z"},"tr":{"title":"Google DeepMind, Gemini için ajan benzeri video anlama özelliğini tanıttı","summary":"Google DeepMind, Gemini 3.7 Flash, 3.6 Flash ve 3.5 Flash-Lite için ajan tarzı video anlayışını tanıttı; model video parçalarını dinamik olarak tarıyor ve sabit kare hızında işleme karşılaştırıldığında token tüketimi en fazla %88, maliyet en fazla %66 azalıyor ve doğruluk en fazla %7 artıyor.","category":"行业动态","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind, Gemini için ajan benzeri video anlama özelliğini tanıttı - Aioga AI Haberleri","description":"Google DeepMind, Gemini 3.7 Flash, 3.6 Flash ve 3.5 Flash-Lite için ajan tarzı video anlayışını tanıttı; model video parçalarını dinamik olarak tarıyor ve sabit kare hızında işleme...","url":"https://www.aioga.com/tr/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:06:58.925Z"},"vi":{"title":"Google DeepMind ra mắt chức năng hiểu video theo dạng agentic cho Gemini","summary":"Google DeepMind đã ra mắt khả năng hiểu video theo tác nhân cho Gemini 3.7 Flash, 3.6 Flash và 3.5 Flash-Lite. Mô hình quét động các đoạn video, so với xử lý theo khung hình cố định thì lượng token tiêu thụ giảm tối đa 88%, chi phí giảm tối đa 66%, độ chính xác tăng tối đa 7%.","category":"行业动态","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind ra mắt chức năng hiểu video theo dạng agentic cho Gemini - Tin tức AI Aioga","description":"Google DeepMind đã ra mắt khả năng hiểu video theo tác nhân cho Gemini 3.7 Flash, 3.6 Flash và 3.5 Flash-Lite. Mô hình quét động các đoạn video, so với xử lý theo khung hình cố địn...","url":"https://www.aioga.com/vi/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:06:57.161Z"},"id":{"title":"Google DeepMind meluncurkan fitur pemahaman video agentik untuk Gemini","summary":"Google DeepMind meluncurkan pemahaman video agen untuk Gemini 3.7 Flash, 3.6 Flash, dan 3.5 Flash-Lite, di mana model memindai potongan video secara dinamis, dibandingkan dengan pemrosesan token dengan frame rate tetap, konsumsi token berkurang hingga 88%, biaya berkurang hingga 66%, dan akurasi meningkat hingga 7%.","category":"行业动态","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind meluncurkan fitur pemahaman video agentik untuk Gemini - Berita AI Aioga","description":"Google DeepMind meluncurkan pemahaman video agen untuk Gemini 3.7 Flash, 3.6 Flash, dan 3.5 Flash-Lite, di mana model memindai potongan video secara dinamis, dibandingkan dengan pe...","url":"https://www.aioga.com/id/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:07:34.698Z"},"th":{"title":"Google DeepMind เปิดตัวฟีเจอร์การเข้าใจวิดีโอแบบเอเยนต์สำหรับ Gemini","summary":"Google DeepMind เปิดตัวการเข้าใจวิดีโอแบบตัวแทนสำหรับ Gemini 3.7 Flash, 3.6 Flash และ 3.5 Flash-Lite โมเดลสแกนชิ้นส่วนวิดีโอแบบไดนามิก เมื่อเทียบกับการประมวลผลที่อัตราเฟรมคงที่ การใช้โทเค็นลดลงสูงสุด 88%, ต้นทุนลดลงสูงสุด 66%, และความแม่นยำเพิ่มขึ้นสูงสุด 7%","category":"行业动态","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind เปิดตัวฟีเจอร์การเข้าใจวิดีโอแบบเอเยนต์สำหรับ Gemini - ข่าว AI Aioga","description":"Google DeepMind เปิดตัวการเข้าใจวิดีโอแบบตัวแทนสำหรับ Gemini 3.7 Flash, 3.6 Flash และ 3.5 Flash-Lite โมเดลสแกนชิ้นส่วนวิดีโอแบบไดนามิก เมื่อเทียบกับการประมวลผลที่อัตราเฟรมคงที่ การ...","url":"https://www.aioga.com/th/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:07:43.327Z"},"pl":{"title":"Google DeepMind wprowadza funkcję agentic do rozumienia wideo dla Gemini","summary":"Google DeepMind wprowadził agentic video understanding dla Gemini 3.7 Flash, 3.6 Flash i 3.5 Flash-Lite, model dynamicznie skanuje fragmenty wideo, w porównaniu z przetwarzaniem o stałej liczbie klatek zużycie tokenów zostało zmniejszone maksymalnie o 88%, koszty zmniejszone maksymalnie o 66%, a dokładność zwiększona maksymalnie o 7%.","category":"行业动态","source":"Google DeepMind：Blog（RSS）","aggregationSource":"Google DeepMind：Blog（RSS）","pageTitle":"Google DeepMind wprowadza funkcję agentic do rozumienia wideo dla Gemini - Aioga Wiadomości AI","description":"Google DeepMind wprowadził agentic video understanding dla Gemini 3.7 Flash, 3.6 Flash i 3.5 Flash-Lite, model dynamicznie skanuje fragmenty wideo, w porównaniu z przetwarzaniem o...","url":"https://www.aioga.com/pl/news/cmtiy0v6k01dproel63pg10os/","contentTranslated":true,"sourceHash":"af7fb3aa6a64e23b","translatedAt":"2026-09-01T18:08:27.677Z"}},"evidenceTier":"verified-news","reviewStatus":"editorial-selected","indexable":true,"editorialCover":"/page-visuals/topic-timeline.png"}}