内容来源：https://blog.google/technology/ai/google-ai-updates-november-2025/

内容总结：

谷歌发布11月人工智能进展：Gemini 3开启智能新纪元，多领域应用全面升级

谷歌近日发布11月人工智能领域最新进展，宣布其研发的Gemini 3模型正式开启“智能新纪元”。该模型在多模态理解、智能代理与创意编程方面实现突破，现已在全球近120个国家和地区的谷歌搜索“AI模式”中向订阅用户开放，提供动态可视化、交互工具等增强搜索体验。

同期推出的Nano Banana Pro图像生成模型基于Gemini 3构建，可生成工作室级高保真视觉内容。全新智能开发平台Google Antigravity进入公开预览阶段，支持开发者部署自主规划复杂任务的AI代理。

在应用生态方面，谷歌地图正集成Gemini实现语音交互导航，支持地标指引与路线规划；Android Auto车载系统也将引入Gemini，提供自然语言操控功能。游戏AI助手SIMA 2通过融合Gemini技术，实现从指令执行到自主进化的跨越。

科研领域，WeatherNext 2天气预测模型将预报速度提升8倍，最高支持1小时精度预报。谷歌同时回顾了AlphaFold蛋白质结构预测模型问世五周年，强调其推动生物科研的持续价值。

为提升用户体验，谷歌搜索新增AI旅行规划工具Canvas，可一键生成定制行程；购物功能引入智能比价与代理结算服务。公司还宣布投入3000万美元支持AI教育项目，并推出NotebookLM应用的闪卡与测验生成功能。

在基础设施布局方面，谷歌宣布将在得克萨斯州投入400亿美元建设AI与云计算设施，作为2025年全球投资计划的重要部分。该计划涵盖欧美、亚非等多地区，并配套推出美国电工培训及学徒岗位创造计划。

中文翻译：

我们于十一月发布的最新人工智能资讯
二十余年来，我们持续投入机器学习与人工智能的研究、工具及基础设施建设，致力于打造让更多人日常生活更美好的产品。谷歌各团队正积极探索人工智能在医疗、危机应对、教育等广泛领域的应用潜能。为及时同步我们的进展，我们将定期汇总谷歌最新人工智能动态。
以下带您回顾十一月部分重要发布。

十一月标志着智能新时代的开启：随着双子座3的发布，我们正提升每个人学习、规划与高效完成任务的能力。
从基于强大模型的"氛围编程"，到通过Nano Banana Pro生成专业级视觉作品，再到全新谷歌反重力平台中的自动化工作流——灵感到现实的距离从未如此之近。无论您是构建复杂智能体的开发者，还是通过画布AI模式规划假期的旅行者，这些更新让人工智能成为主动协作的伙伴，随时助您高效达成目标。

我们发布了开启智能新时代的双子座3。双子座3旨在让任何创意落地生根，它代表着我们在拓展智能边界、打造代理体验与个性化服务方面迈出的新步伐，力求让人工智能真正惠及每个人。作为全球顶尖的多模态理解模型，双子座3是我们迄今最强大的代理与氛围编程模型。对开发者而言，双子座3 Pro在主要人工智能基准测试中全面超越前代版本。双子座3升级的智能与全新功能现已登陆双子座应用，您可通过我们的专题页面查阅所有相关发布。

我们将双子座3融入谷歌搜索，打造迄今最智能的搜索体验。双子座3先进的推理能力现可通过AI模式在谷歌搜索中使用——这是我们首次在发布当日便将双子座模型集成至搜索功能。该模型能理解深度与细微差异，并通过动态视觉布局、交互工具及针对查询定制的模拟功能，开启搜索新体验。全球近120个国家和地区的谷歌AI Pro与Ultra订阅用户现可在AI模式中通过模型下拉菜单选择"使用3 Pro思考"来体验该功能。

我们推出了基于双子座3打造的Nano Banana Pro。这款全新图像生成编辑模型突破即兴创作范畴，开启高保真影棚级视觉新时代。用户既可选择原版Nano Banana进行趣味快速编辑，也可选用Pro版本获得能处理高质量复杂任务的强大创作伙伴。我们同步分享了七条技巧助您充分发挥其潜能。

我们推出了全新代理开发平台谷歌反重力。该平台为开发者提供超越简单编辑的人工智能编程体验，通过首创的代理优先界面部署可自主规划、执行与验证复杂任务的智能体。我们期待反重力平台让每个创意都能腾飞落地。该平台现已开放公开预览。

我们宣布谷歌地图因双子座而更智能。借助双子座技术，您将很快在谷歌地图中体验首款免提对话式驾驶功能——仅需语音指令即可搜索地点、上报路况、获取沿途建议等。新增地标导航功能将提供清晰指引，除了"500英尺后右转"这类提示，还会结合"泰式暹罗餐厅后右转"等实用地标进行引导。地标导航功能正于美国安卓与iOS平台逐步推出，集成双子座的导航功能将在双子座已上线地区全面铺开。

我们宣布双子座开始登陆Android Auto。目前全球已有超过2.5亿辆汽车搭载Android Auto系统，双子座的加入将进一步提升驾乘体验。驾驶途中您可使用自然语言添加途经点、发送消息、查阅邮件、创建歌单甚至进行创意构思。只需确保手机已安装双子座应用，并在车载屏幕上寻找功能提示即可使用。

我们推出SIMA 2，向通用人工智能迈出重要一步。这是我们打造通用实用人工智能代理进程中的重要里程碑。通过融合双子座的先进能力，SIMA 2正从指令执行者转型为互动游戏伙伴。如今它不仅能遵循虚拟世界中的自然语言指令，还能自主思考目标、与用户对话并持续进化，标志着我们在机器人技术与人工智能实体化方向迈出关键步伐。

我们发布WeatherNext 2：迄今最先进的天气预报模型。该模型生成预报的速度提升8倍，最高支持1小时级精度预报。我们已运用这项突破性技术协助气象机构进行决策。

我们纪念AlphaFold破解蛋白质折叠问题五周年，并聚焦其持续产生的深远影响。五年前AlphaFold 2解决了蛋白质结构预测难题，这项工作的重大科学与社会价值于2024年荣获诺贝尔化学奖。十一月我们回顾了AlphaFold如何开辟生物学研究新路径，首次有力证明了人工智能可成为推动科学进步的强大工具。

我们宣布在搜索中推出人工智能旅行规划新方式。搜索中的新人工智能功能可协助构建完美行程——从获取超值优惠到将计划转为实际预订。只需在AI模式中尝试画布工具：描述理想旅程与所需建议，点击"创建画布"即可生成专属旅行方案。

我们在搜索与双子座中新增假日购物人工智能功能。本次购物功能重磅升级让您能通过对话式人工智能与代理人工智能简化假日采购。使用搜索AI模式购物时，描述需求即可获得智能整合的响应，包含丰富视觉展示、价格、评价、库存等信息。新增的人工智能代理还能致电商店确认库存，并在价格合适时通过代理结算功能自动向合作商户完成购买。

我们宣布对人工智能学习与教育的新承诺。在伦敦举办的人工智能学习论坛上，我们宣布新增3000万美元学习基金。这场汇聚教育科技专家的活动延续了我们近期通过人工智能提升学习成效的探索。我们还发布了助力日常学习的热门新功能，包括在NotebookLM应用中直接创建记忆卡与测验。

我们分享使用双子座Live的核心技巧。最新更新让您能与人工智能进行更自然的双向对话，我们的技巧演示了如何通过个性化学习设置（特别是复杂科目）充分发挥这些功能。您现在可调节语音速度按自身节奏学习，还能练习外语、模拟面试，甚至通过趣味口音让对话更生动。

我们宣布在德克萨斯州新增400亿美元人工智能与云基础设施投资。谷歌与Alphabet首席执行官桑达尔·皮查伊与德州州长格雷格·阿博特在米德洛锡安市共同宣布了这一消息。此项投资标志着我们2025年人工智能投资计划的收官之作，旨在创造经济机遇、推动科学突破、惠及全球大众。投资范围覆盖美洲、欧洲、非洲及亚太地区，同步启动的美国人才计划将培训10万名电气工人并新增3万个学徒岗位。

英文来源：

The latest AI news we announced in November
For more than 20 years, we’ve invested in machine learning and AI research, tools and infrastructure to build products that make everyday life better for more people. Teams across Google are working on ways to unlock AI’s benefits in fields as wide-ranging as healthcare, crisis response and education. To keep you posted on our progress, we're doing a regular roundup of Google's most recent AI news.
Here’s a look back at some of our AI announcements from November.
November ushered in a new era of intelligence: with the launch of Gemini 3, we're improving everyone's ability to learn, plan or just get things done.
From "vibe coding" with our powerful models, to generating professional grade visuals with Nano Banana Pro, to the autonomous workflows in our new Google Antigravity platform, the distance between a spark of imagination and reality has never been shorter. So, whether you’re a developer building complex agents or a traveler planning for the upcoming holidays with Canvas in AI Mode, these updates turn AI into a proactive partner ready to help you get things done, all season long.
We released Gemini 3, AI for a new era of intelligence. Gemini 3 is built to bring any idea to life. It represents the next step in our work to push the frontiers of intelligence, agentic experiences and personalization — so that AI is truly helpful for everyone. Gemini 3 is the best model in the world for multimodal understanding and it's our most powerful agentic and vibe coding model to date. For developers, Gemini 3 Pro outperforms previous versions across major AI benchmarks. Gemini 3’s upgraded smarts and new capabilities are now available in the Gemini app, and you can review our hub with all the Gemini 3 announcements.
We made Gemini 3 available in Google Search for our most intelligent search yet. Gemini 3’s state-of-the-art reasoning is now available in Google Search, starting with AI Mode — marking the first time we brought a Gemini model to Search on day one. Gemini 3 grasps depth and nuance, and unlocks new experiences in Search with dynamic visual layouts, interactive tools and simulations tailored specifically for your query. Google AI Pro and Ultra subscribers in nearly 120 countries and territories in English can use Gemini 3 Pro by selecting “Thinking with 3 Pro” from the model drop-down menu in AI Mode.
We unveiled Nano Banana Pro, built on Gemini 3. Nano Banana Pro, our newest image generation and editing model, is built on Gemini 3 and moves beyond spontaneous art into an era of high-fidelity, studio-quality visuals. You now have the choice between the original Nano Banana for fun, fast editing, or Pro for an even more powerful creative partner capable of handling complex tasks demanding the highest quality. To help get you started, we shared seven tips to get the most out of Nano Banana Pro.
We introduced Google Antigravity, a new agentic development platform. Antigravity is a platform designed to give developers an AI-powered coding experience that goes beyond simple editing. It delivers a new agent-first interface for deploying agents that autonomously plan, execute and verify complex tasks. Our vision for Antigravity is to enable anyone with an idea to experience liftoff and build that idea into reality. You can try Antigravity available for yourself today in public preview.
We announced that Google Maps is getting smarter with Gemini. With the help of Gemini, you will soon have the first, hands-free, conversational driving experience in Google Maps that allows you to find places, report traffic, ask for suggestions along your route and more using just your voice. Plus, new landmark-based navigation will give you clear directions, so in addition to hearing “turn right in 500 feet,” you’ll also get directions based on helpful landmarks like “turn right after the Thai Siam Restaurant.” Landmark-based navigation is rolling out now on Android and iOS in the U.S, and Gemini in navigation on Google Maps is rolling out everywhere Gemini is available.
We announced that Gemini has started rolling out in Android Auto. Android Auto is already available in over 250 million cars on the road, and Gemini is now coming along for the ride to make life on the road even better. You’ll be able to use natural language to add stops, send messages, access emails, create playlists and even brainstorm ideas while driving. Just make sure you have the Gemini app on your phone and look for the tooltip on your car display.
We introduced SIMA 2, a significant step toward Artificial General Intelligence (AGI). SIMA 2 is a major milestone in our work to create general and helpful AI agents. By integrating the advanced capabilities of Gemini into SIMA 2, the model is evolving from an instruction-follower into an interactive gaming companion. Now, SIMA 2 can follow human-language instructions in virtual worlds and also think about its goals, converse with users and improve itself over time, marking an important step in the direction of robotics and AI-embodiment in general.
We released WeatherNext 2: our most advanced weather forecasting model. WeatherNext 2 can generate forecasts 8x faster and with resolution up to 1-hour, and we’re already using this breakthrough technology to support weather agencies in making decisions.
We marked the 5-year anniversary of AlphaFold cracking the protein folding problem, and spotlighted its ongoing impact. Five years ago AlphaFold 2 solved the protein structure prediction problem. The profound scientific and societal value of this work was recognized in 2024 with the Nobel Prize in Chemistry. In November, we looked back at how AlphaFold has unlocked new avenues of biological research and provided our first major proof point that AI can be a powerful tool to advance science.
We announced new ways to plan travel with AI in Search. Our new AI features in Search help you build the perfect itinerary, from snagging a great deal to turning your plans into actual bookings. To get started with planning, try the Canvas tool in AI Mode: Describe the kind of trip you want and what recommendations you need. Then hit "Create Canvas," and watch your custom travel plan come together.
We added new AI shopping features in Search and Gemini to help with the holidays. Our biggest upgrade to shopping means you can now use conversational AI and agentic AI to take the hard work out of your holiday shopping. Using AI Mode in Search to shop, you can now describe what you’re looking for and get an intelligently organized response that brings together rich visuals, price, reviews, inventory info, and more. Plus, new AI agents can now call stores to check stock and use agentic checkout to buy items automatically from eligible merchants when the price is right.
We announced new commitments to AI learning and education. At our AI for Learning Forum in London, we announced $30 million in new funding for learning. The event brought together experts in education and technology and continued our recent work to develop AI in a way that improves learning outcomes. We also released popular new features to help with daily studying, including the ability to create flashcards and quizzes right in the NotebookLM app.
We shared essential tips for using Gemini Live. Gemini Live’s latest updates allow you to have more natural, two-way conversations with AI. Our tips show how to make the most of these updates by tailoring your learning, especially for complex subjects. You can now adjust Gemini's speech speed to learn at your own pace and improve accessibility. You can also practice a new language, rehearse for a job interview or even liven your conversation up with a fun accent.
We announced a new $40 billion investment in Texas for AI and cloud infrastructure. CEO of Google and Alphabet Sundar Pichai and Texas Governor Greg Abbott made the announcement at an event in Midlothian, TX. This latest announcement represented the capstone of our 2025 push to make AI investments that unlock economic opportunity, advance scientific breakthroughs and create opportunities that benefit everyone. It includes major investments across America, as well as Europe, Africa and the Asia-Pacific region — alongside a critical US workforce initiative to train 100,000 electrical workers and create 30,000 new apprentices.