Google

Google develops its AI models at Google DeepMind, which brings together two AI labs, Google Brain and DeepMind, under the leadership of Demis Hassabis. Google DeepMind states its mission as building AI responsibly to benefit humanity. Google itself was founded in 1998 and is based in Mountain View, California.

Gemini is Google's main family of models; the current generation in the Gemini API documentation is Gemini 4 Argon. Around it sit specialised families: Nano Banana for image generation and editing (currently Nano Banana 2.1), Veo for video with audio (Veo 3.1), Lyria for music (Lyria 3.5) and Gemma, a family of open-weight models (Gemma 4). AlphaFold predicts the structures of proteins.

Google ships these models in many products. The Google Gemini app is its everyday assistant, and Gemini Notebook (formerly NotebookLM) works with your own sources. Google AI Studio and the Gemini Enterprise Agent Platform on Google Cloud (formerly Vertex AI) give developers and companies access to the models. Google Antigravity is an agent-first development platform, Stitch turns prompts and sketches into interface designs, and Opal builds small AI apps from natural language.

This page lists the official Google AI products in the Best-AI.org catalog and the model families behind them, each with its source and the date it was checked.