Gemini

Gemini is Google's AI assistant, built on a family of models that has moved well past basic Q&A. The current flagship for everyday tasks is Gemini 3.5 Flash, optimized for agentic and coding work. For speed and cost efficiency there's Gemini 3.6 Flash and 3.5 Flash-Lite. The newest addition, Gemini 3.8 Flash (GA as of September 2026), is built specifically for long-horizon software engineering, autonomous agents, and complex enterprise workflows.

Beyond text, Gemini now handles a wide range of modalities. Gemini Omni Flash generates and edits video conversationally, supporting 360p through 4K output, video extension, and two-image interpolation. Gemini 3.5 Transcribe offers high-accuracy speech-to-text across 85+ languages with speaker diarization and word-level timestamps; a Live variant handles real-time streaming over WebSockets. Native image models (Gemini 3.1 Flash Image, Gemini 3 Pro Image) support video-to-image generation, including thumbnails from YouTube URLs. Music generation is handled by Lyria 3.5, which produces full-length songs with structural control at 44.1 kHz stereo.

For developers, Gemini now supports Managed Agents — stateful, autonomous agents running in Google-hosted Linux sandboxes that can plan, write and execute code, manage files, and browse the web. The Antigravity Agent is the general-purpose version of this. Agentic video understanding (available on 3.7, 3.6, and 3.5 Flash-Lite) dynamically navigates video timelines on demand, using up to 88% fewer tokens than static processing. Computer Use is in public preview on Gemini 3.5 Flash, with support for browser, mobile, and desktop environments.