Gemini is Google's leading family of AI models. It's the flagship of Google DeepMind and the model that powers all of Google's AI products, from Search Generative Experience to Workspace, Android and Bard.
The main models
- Gemini 2.5 Pro — the most advanced model, with deep reasoning and a 2M token context window (the largest in the industry).
- Gemini 2.5 Flash — a fast, cheap version, optimized for scale.
- Gemini Nano — a small model that runs directly on the device (Pixel, Samsung), with no internet connection needed.
- Gemini 3 Pro Image (Nano Banana Pro) — one of the most advanced image generation models in the world, with photorealistic quality and support for editing existing images.
Key capabilities
- Multimodal native — works with text, images, audio and video in a single model, from the first stage.
- Tool use — can call functions, search Google, run Python code and integrate information sources.
- Long context — 2M tokens allow analysis of hours-long videos, entire books and huge codebases.
- Live API — real-time voice conversation with image understanding through the camera.
Gemini is available via Google AI Studio (free for developers), Vertex AI (Enterprise), the Gemini App (gemini.google.com), and integrated into Workspace (Docs, Gmail, Sheets) and Android Pixel devices.
On this site: the OG image, the missing Top Engineering project image, and other generated visuals on the site were created with Gemini 3 Pro Image ("Nano Banana Pro") via the API.
—BeeUONO
