Google Gemini

Google Gemini runs on a family of models including Gemini 3 Flash for fast everyday tasks, Gemini 3 Pro for advanced reasoning and coding, and Gemini 3.1 Pro for the most complex problem-solving. Deep Think is a specialized reasoning mode that explores multiple hypotheses simultaneously for science and math challenges.

Deep Research automatically browses and analyzes hundreds of websites in real-time using the Gemini 3 Pro model. It creates a research plan, executes searches, synthesizes findings, and delivers a multi-page report in minutes. Users can upload their own files as additional sources and transform reports into interactive content via Canvas.

The free tier provides a standard context window, while AI Plus subscribers get a 128K token context window. AI Pro and AI Ultra subscribers unlock a 1-million token context window, allowing them to process extremely large documents, codebases, and multi-turn conversations in a single session.

Yes, Gemini offers Veo 3.1 for generating 8-second videos with synchronized audio including dialogue, background music, and sound effects. AI Plus and Pro users access Veo 3.1 Lite, while AI Ultra users get access to the full Veo 3.1 model. Music generation is available through the Lyria 3 model for creating custom soundtracks from text prompts.

Personal Intelligence connects to Gmail, Photos, YouTube, and Search to provide context-aware responses, but it is off by default. Users must explicitly opt in, choose which apps to connect, and can revoke access at any time. Data connections are managed through granular settings, and Temporary Chat mode ensures certain conversations are not saved or used for model training.