At its I/O 2026 developer conference on May 19, Google announced a significant expansion of its Gemini product line, headlined by a new multimodal model called Gemini Omni and a persistent cloud-based agent called Gemini Spark. The announcements were made alongside the launch of the Gemini 3.5 model family and a redesigned Gemini application interface.
Gemini Omni and the 3.5 model family
<cite index="11-11">Google described Gemini Omni as a new model that can transform text, image, and video prompts into cinematic, high-quality video outputs.</cite> According to Google Cloud documentation, <cite index="4-14">Omni is positioned as a leap forward in world understanding, multimodality, and editing, letting users generate any output from any input, starting with video.</cite> <cite index="3-14">Gemini Omni Flash accepts image, audio, video, and text input and outputs video grounded in real-world knowledge that can be easily edited.</cite> <cite index="13-11">Gemini Omni began rolling out to Google AI Plus, Pro, and Ultra subscribers worldwide on the day of the keynote.</cite>
The Omni release sits within the broader Gemini 3.5 series. <cite index="2-8,2-10">Google launched Gemini 3.5 Flash, the first in the new model series combining frontier intelligence with action, generally available via Google Antigravity, the Gemini API in Google AI Studio, and Android Studio.</cite> <cite index="2-13">Google reported that 3.5 Flash outperforms Gemini 3.1 Pro on coding and agentic benchmarks including Terminal-Bench 2.1 (76.2%), GDPval-AA (1656 Elo), and MCP Atlas (83.6%).</cite>
Gemini Spark agent
<cite index="11-13,11-14">Google introduced Gemini Spark as a 24/7 personal AI agent intended to transform Gemini from an assistant that answers questions into an active partner that does work on the user's behalf.</cite> <cite index="11-15,11-16,11-17">Spark runs on Gemini 3.5 and uses the Antigravity harness, integrates with Workspace tools including Gmail, Docs, and Slides, and continues operating in the background even when a user closes a laptop or locks a phone.</cite>
Spark also supports the Model Context Protocol (MCP), an open standard for connecting agents to external services. <cite index="9-19,9-20">At launch, Spark ships with MCP connections to Canva, OpenTable, and Instacart, with Adobe, Samsung, Spotify, CapCut, GitHub, Notion, and Slack confirmed for summer 2026.</cite> <cite index="11-27">Google said Spark would roll out to trusted testers in the week of the announcement and as a Beta for U.S. Google AI Ultra subscribers the following week.</cite>
Scale and pricing
<cite index="11-7,11-8">Google reported that Gemini served 400 million users at last year's I/O and now reaches more than 900 million people across 230 countries and more than 70 languages each month.</cite> Alongside the product launches, Google restructured its subscription pricing. <cite index="14-17,14-18">The AI Ultra tier was reduced from $250 to $100 per month, with the $100 plan including five times the usage limits of the $20 AI Pro plan, 20 terabytes of cloud storage, YouTube Premium, and beta access to Spark.</cite>
Interface and additional agents
<cite index="2-28">Google introduced a new design language called Neural Expressive, with fluid animations, vibrant colors, new typography, and haptic feedback.</cite> A second new agent, Daily Brief, was also announced. <cite index="10-5">Daily Brief, a morning-digest agent that summarizes inbox and calendar items, launched on the day of the keynote for AI Plus, Pro, and Ultra subscribers in the United States.</cite>
Google framed the updates as part of a broader shift from reactive chatbots to persistent agentic systems, supported by continued capital expenditure in custom silicon. <cite index="1-31,1-32">The company said capital expenditure, which was $31 billion annually in 2022, is expected to reach approximately $180 to $190 billion this year.</cite>