Google 推出 Gemini 3，强调其在推理、视频、多模态与代码生成上的跃升，并直接接入搜索与 Gemini 应用。数据侧的规模效应突出：Gemini 应用月活 6.5 亿人；AI Overviews 月活 20 亿人；开发者生态达 1,300 万人。Google 称 Gemini 3 在 LMArena 等基准上优于 GPT-5，并在长程规划与分解推理任务上取得领先。自然语言查询同比增长达“双位数”，视觉搜索量激增 70%，AI Overviews 带动搜索查询量在最新季度增长 10%。付费定位方面，公司将 Gemini 3 下放至 Google AI Plus（$19.99/月）与 Google AI Pro（$249.899/月）。
在商业化上，Google 试图以 Gemini 3 稳固搜索收入：用户检索如“三体问题”等内容时，将自动生成可交互可视化；Google 亦推进与 Apple 的导入谈判，力图将 Gemini 融入 Siri。辅助产品线扩展包括 NotebookLM、AI Studio，以及全新的编程工具 Antigravity。此外，Google 强调其全栈优势：自研芯片、数据中心规模与巨量用户交互提供持续训练信号，使模型能快速定位薄弱知识区域并迭代。
尽管市场存在“AI 泡沫”担忧，Google 认为其基本盘（地图、Gmail、搜索）能在下行情境吸收风险，而上行情境下其产品组合与研究深度可放大回报。Gemini 3 被定位为未来五到十年迈向 AGI 的平台层模型，仍需一到两次关键突破方能达到公司口中的“完全 AGI”。
Google launched Gemini 3 with major gains in reasoning, video generation, multimodal processing, and code synthesis, integrating it directly into Search and the Gemini app. Usage-scale metrics are substantial: 650 million monthly users for the Gemini app, 2 billion monthly users for AI Overviews, and 13 million developers building on Google models. Google claims Gemini 3 outperforms GPT-5 on benchmarks like LMArena, with superior long-horizon planning and structured reasoning. Natural-language queries have risen by “double digits” year over year, visual search is up 70%, and AI Overviews drove a 10% increase in search queries in the latest quarter. The model will reach Google AI Plus ($19.99/month) and AI Pro ($249.899/month).
Commercially, Google aims to anchor search revenue with Gemini 3: queries such as those about the three-body problem now auto-generate interactive visualizations. The company is also negotiating to bring Gemini to Apple’s Siri. Complementary products—NotebookLM, AI Studio, and the new Antigravity coding tool—expand the ecosystem. Google stresses full-stack advantages: in-house silicon, hyperscale data centers, and massive user feedback loops that reveal model weaknesses and accelerate refinement.
Despite growing “AI bubble” concerns, Google argues its core products (Maps, Gmail, Search) buffer downside scenarios, while its research depth amplifies upside potential. Gemini 3 is positioned as a platform for the next 5–10 years of progress toward AGI, with one or two further breakthroughs deemed necessary to achieve what the company considers full AGI.