Gemini
Google's family of natively multimodal models, available in app, API, Workspace, and Android.
- Sizes: Nano, Flash, Pro, Ultra.
- Native handling of text, images, audio, video.
- Long context (1M-2M tokens) and Google Search integration.
When to use: For document-grounded work or when Google ecosystem integration matters.
Example: Gemini 1.5 Pro ingests a 1-hour video, transcribes it, and answers timeline questions in one API call.