FROM THE FRONTIER
Everything you need to know about Gemini 3
Source: Google.
Gemini 3 is** now live** in AI Mode, the Gemini app, and AI Studio. These are the key features you should know about:
- A major leap in multimodal reasoning: Gemini 3 outperforms leading AI models across every major benchmark (except coding). It also processes information seamlessly across text, images, video, audio, and code.
- Generates concise, direct responses: Google designed Gemini 3 to deliver straight-to-the-point answers that “tell you what you need to hear, not just what you want to hear.”
- Offers better results with less precise prompting: Gemini 3 excels at understanding the intent behind your prompts and delivering relevant information even when your prompts are brief or vague.
- AI Mode on Search offers custom visualizations: AI Mode now generates custom images, tables, interactive tools, and “on-the-fly” simulations to answer queries visually. See how to use it.
Google also released 3 new Gemini-powered tools: - Gemini Agent: A tool that takes action across Google’s apps to help with tasks like organizing your inbox or booking services. Available for Google AI Ultra subscribers. See Agent in action.
- Deep Think mode: An advanced reasoning model that tackles your most complex queries by enhancing Gemini 3’s capabilities.
- Google Antigravity: An agent-first coding tool built for Gemini 3 that supports multiple agents with direct access to your editor, terminal, and browser.