At a glance: Google launched Gemini 3.5 Flash immediately at I/O 2026 and introduced the new Omni model family for video processing. With advanced agent capabilities through Antigravity 2.0 and processing 3.2 quadrillion tokens monthly, Google demonstrates its technological leadership in the AI industry.
Google presented a comprehensive realignment of the Gemini platform at its annual developer conference I/O 2026. With Gemini 3.5 Flash, the new Omni model for video processing, and the expanded Antigravity agent suite, the company demonstrated its leadership position in AI technology and latency performance.
Google repositioned Gemini as a central platform for consumer and developer applications at I/O 2026. The keynote, whose highlights are documented in a 30-minute recap, presented three core technical announcements.
The centerpiece of the announcements is Gemini 3.5 Flash, now generally available and positioned as the strongest model for agent-driven tasks and code generation. It features an extended context of one million tokens, a maximum output length of 65,000 tokens, and four different “thinking” levels ranging from minimal to highly complex. The model retains its reasoning across multiple conversation turns, making it particularly suitable for iterative development processes.
Alongside 3.5 Flash, Google presented the new Omni model family, which combines video generation and editing with Gemini’s reasoning capabilities for the first time. Omni Flash enables processing of text, image, video, and audio inputs and creates video edits and new video content from them, available initially in Gemini, Google Photos Flow, and YouTube Shorts with upcoming API access.
The Antigravity 2.0 suite extends agent capabilities through desktop, CLI, and SDK interfaces. With the new Spark feature, AI agents can operate in the background on cloud VMs. Additionally, Google integrated the new models with search-native generative UI and numerous integrations into Workspace, Commerce, and Media applications.
The figures underscore Google’s scaling: the company now processes 3.2 quadrillion tokens monthly—a sevenfold increase from 480 trillion in the prior year. Gemini reached 900 million monthly active users and is available in over 230 countries and 70 languages. Gemini 3.5 Flash is immediately available on all platforms, including the Gemini App, Search AI Mode, the Gemini API, AI Studio, Antigravity, and Android Studio.