Google launches Gemini 3.5 Flash and Gemini Spark at I/O 2026. The new 24/7 AI agent and cinematic video model redefine the digital experience.
Google unveiled a massive overhaul of its artificial intelligence ecosystem at the Shoreline Amphitheatre today. The tech giant introduced Gemini 3.5 Flash, a high-speed model designed for complex agentic tasks. CEO Sundar Pichai also announced Gemini Spark, a personal agent that works autonomously in the cloud. This strategic shift moves Google beyond simple chatbots into the era of proactive digital partners.
Google slashes AI Ultra prices
Consequently, the entry price for Google’s premium AI services has reached a new low. The company officially cut the cost of its AI Ultra plan from $250 to just $100 per month. This move targets power users who require higher computational limits for heavy coding and creative workflows. Forbes notes that this aggressive pricing strategy aims to capture the enterprise market from rivals.
Furthermore, the subscription now includes five times the usage limits of the standard Pro tier. Users also gain access to 20 terabytes of cloud storage and premium YouTube features as part of the bundle. This pricing restructure coincides with a new “compute-used” metering system for all Gemini applications. Much like the CBN’s move to deepen market liquidity, Google is seeking to stabilize its user base through transparency and value.
Gemini Spark: The agent that never sleeps
Additionally, the event featured the debut of Gemini Spark, a 24/7 personal AI assistant. Unlike traditional chatbots, Spark runs on virtual machines in the Google Cloud to perform tasks while users are offline. It can organize emails, draft documents, and manage complex schedules across Google Workspace apps. According to 9to5Google, Spark represents a fundamental shift toward agentic AI that acts on your behalf.
Moreover, the agent integrates deeply with a new macOS desktop application launching this summer. This software allows Spark to handle local files and automate repetitive workflows directly on a user’s machine. It utilizes a new design language called Neural Expressive to provide fluid, haptic-enabled interactions. This interface uses vibrant colors and interactive timelines to make AI responses feel more human and engaging.
Cinematic video with Gemini Omni
Meanwhile, Google DeepMind introduced Gemini Omni to tackle the growing demand for high-quality generative media. This multimodal model transforms text, images, and audio into cinematic, high-fidelity video outputs. It effectively fills the market gap left after OpenAI discontinued its Sora video platform last month. Omni allows for conversational video editing, enabling users to swap backgrounds or add objects through simple voice prompts.
Specifically, the Omni Flash variant is rolling out today to subscribers of all premium tiers. It uses advanced “world models” to simulate real-world physics and lighting for more realistic results. Experts at TechRadar claim that Gemini 3.5 Flash is now four times faster than previous flagship models. Developers can also access these tools through the newly updated Antigravity 2.0 platform starting this week.
Google’s latest updates signal a massive push toward making AI a constant, invisible partner in daily life. By lowering costs and increasing autonomy, the company is positioning Gemini as an essential utility for the modern age. The global rollout of these features begins today across Android, iOS, and the web.
