Google used its annual I/O developer conference in May 2026 to announce Gemini 3.5 Flash, a faster and cheaper version of its flagship AI model, alongside Gemini Spark, a new general-purpose AI agent designed to reason across connected applications and take action on behalf of users.
Gemini 3.5 Flash reflects Google's strategy of deploying capable models at cost points that allow them to run across products used by hundreds of millions of people, rather than reserving frontier capabilities for premium tiers only. The model is intended to power features embedded across Google Search, Workspace, and Android.
Gemini Spark, in beta at the time of the conference, is available first to trusted testers and Google AI Ultra subscribers. The agent can connect to Gmail, Google Drive, Calendar, and third-party apps, allowing it to reason across a user's information environment and surface relevant context or take actions without explicit prompting for each step.
Google also reported that Gemini now has 900 million monthly active users, more than double the approximately 400 million reported at the same time the prior year. The growth reflects both consumer adoption through the Gemini app and developer deployment through Google Cloud's Vertex AI platform.
Executives at Google described the frontier AI race as effectively neck and neck across the leading providers, with companies making different tradeoffs between model cost, response speed, and reasoning capability. Google's I/O announcements landed in the same week as significant product releases from competing platforms.
Source: CNBC -- https://www.cnbc.com/2026/05/19/google-ai-ultra-gemini-spark-omni.html