Gemini

Gemini is Google DeepMind's flagship family of large language models, succeeding the earlier PaLM and Bard models. The family spans multiple sizes and specializations: Gemini 3.1 Pro (the most capable), Gemini Flash (optimized for speed and cost), and Gemini Nano (designed for on-device inference on mobile hardware). All Gemini models are natively multimodal, meaning they can process and generate text, images, audio, and video.

Gemini is deeply integrated across Google's product ecosystem. It powers AI features in Google Search, Gmail, Google Docs, Google Sheets, Google Ads, Google Lens, and Android. The Gemini API and AI Studio provide developer access for building applications on top of the models. Unlike some competitors, Google offers Gemini with context windows up to 1 million tokens, enabling processing of entire codebases, long documents, and extended conversations.

In January 2026, Apple announced a multiyear partnership with Google to use Gemini models and cloud infrastructure to power the next generation of Apple Foundation Models, including a more personalized Siri. This deal positioned Gemini as the intelligence layer underneath both Android and iOS, the two dominant mobile operating systems. Google also develops specialized Gemini variants like Nano Banana (image generation) and continues to release open weight models in the Gemma family for researchers and developers who want to run models locally.