Human
MOUNTAIN VIEW, CA: Gemini 3, the latest artificial-intelligence model from Google, has officially launched on Nov. 18, 2025 and is immediately being embedded into the company's flagship products, including its search engine. Google describes Gemini 3 as its "most intelligent model," citing performance gains in reasoning, coding and multimodal capabilities (text, image, audio and video).
In a major shift, the rollout is designed to integrate the model into user-facing services from day one. Earlier versions of the company's AI models were introduced with delay between announcement and product integration: Gemini 3 marks the first instance where Search will be powered by the new model at launch. One showcased feature, dubbed "Gemini Agent", allows the system to conduct multi-step tasks such as organising email or booking travel, while the redesigned Gemini app can provide interactive responses, visualisations and multimedia content.
Google's launch comes amid intensifying competition with OpenAI and others jockeying to lead in frontier-AI developments. The company says that the new model not only delivers higher technical benchmarks but is positioned to support revenue-generating services immediately, a shift from earlier models largely which were frequently used for research or preview.
Despite the optimism, the broader AI market remains cautious. Analysts warn of a potential "AI bubble" and emphasise that future growth will depend on commercial uptake rather than benchmark scores alone. With Gemini 3 now live, Google aims to transform its search business, workspace tools and app platforms into intelligent assistants, ushering in a new phase of AI-driven consumer technology.
