Google unveiled Gemini 3, its most advanced AI model to date, positioning it as state-of-the-art in multimodal reasoning and long-horizon planning. The company said Gemini 3 Pro leads key benchmarks, adds a Deep Think mode for tougher reasoning tasks, and integrates across Google products on day one, including AI Mode in Search, the Gemini app, AI Studio and Vertex AI. Google also introduced Antigravity, an agent-centric development platform that lets AI agents plan and execute end-to-end software tasks with access to an editor, terminal and browser.
The rollout underscores Google’s push to commercialize agentic AI, with monetization through Pro and Ultra subscriptions and enterprise offerings. Google cited momentum across its ecosystem—2 billion monthly users of AI Overviews, 650 million monthly Gemini app users, 70% of Cloud customers using AI, and 13 million developers building with generative models. The company emphasized expanded safety testing and external assessments as it competes with OpenAI, Anthropic and Meta to define the next phase of AI-driven products.
Related article:
