Google launches dual-tier AI Ultra pricing at I/O 2026, adds Gemini Spark agent and Project Genie expansion

Google is fragmenting its AI subscription tiers into two distinct Ultra plans, priced at $100 and $200 per month, while cutting the previous top-tier cost from $250 to $200. According to the Google AI Blog, the announcement came at the company’s I/O 2026 developer conference and targets different user profiles—the $100 plan aims at developers seeking iterative AI tools, while the $200 plan caters to advanced builders requiring agent deployment and generative world exploration.

The $100 AI Ultra tier for developers and technical workers

The entry-level Ultra plan bundles 5X higher usage caps in Gemini and Google Antigravity than the Pro tier, bundled with Gemini 3.5 Flash for rapid testing and debugging cycles. Google pairs this with 20TB of cloud storage (positioning the plan for teams managing large codebases or datasets) and includes YouTube Premium to enable background content consumption without interruption.

The tier specifically targets engineers and technical leads seeking to compress development workflows without committing to the premium $200 option. According to Google AI Blog, Antigravity—the company’s development framework—receives priority queuing for $100 subscribers, distinguishing it from lower tiers.

Premium $200 plan: price cut plus agent and world-creation features

Google reduced its top-tier monthly cost from $250 to $200, preserving the 20X usage advantage over Pro while adding two experimental capabilities. According to the Google AI Blog, Gemini Spark is a 24/7 agent designed to execute tasks across Google’s product ecosystem on the user’s behalf, rolling out first to trusted testers before moving to beta access for US-based Ultra subscribers. The company also expanded Project Genie access to all eligible $200 subscribers globally (age 18 and older), introducing a new Street View-powered mode for generating novel environments.

Both Ultra tiers receive Gemini Omni, Google’s latest multimodal foundation model, enabling native audio, video, and text generation within a single inference pipeline.

Why This Matters

Google’s dual-tier Ultra strategy signals a shift toward segmentation by velocity: builders needing real-time iteration tools (the $100 audience) versus those piloting agentic workflows and experimentation (the $200 base). The $250-to-$200 price normalization removes a premium-tier anchoring effect, potentially expanding the addressable market for Antigravity and reducing friction for organizations evaluating agent-building platforms. For developers already embedded in Google’s ecosystem (Workspace, Cloud), the bundled YouTube Premium and expanded storage lower switching costs to enterprise AI subscriptions. The staged rollout of Gemini Spark and Project Genie to subscriber subsets creates a conversion funnel from testing (trusted beta) to production use, hedging against undiscovered capability gaps before public launch.