A New Middle Lane in Google Gemini Pricing
Google is reshaping its premium AI lineup with a clearer ladder for heavy users. At Google I/O, the company unveiled a new AI Ultra tier priced at USD 100 (approx. RM460) per month and reduced its former top Ultra subscription from USD 250 (approx. RM1,150) to USD 200 (approx. RM920). This move narrows the gap between the mainstream Google AI Pro plan and the most expensive option, giving developers, technical leads, and advanced creators a less costly entry into Google’s highest-end consumer AI bundle. The AI Ultra tier offers 20 TB of storage, YouTube Premium, and Gemini app plus Google Antigravity allowances that are five times higher than AI Pro. The upper Ultra plan keeps the same features but raises usage ceilings to roughly 20 times Pro, cementing capacity—not features—as the main distinction across Google Gemini pricing tiers.
From Subscriptions to Compute-Based Usage Limits
Alongside new tiers, Google is abandoning simple daily prompt caps in favor of compute-based usage limits. Instead of counting how many queries a user sends, Google now meters the underlying processing power each interaction consumes. Usage refreshes every five hours until a weekly ceiling is reached, with paid top-up credits planned for those who regularly exceed their allowance. This approach aligns cost with actual workload intensity, making it easier for buyers to estimate how coding sprints, chained research, or media-heavy tasks will affect their budget. Previously, plans could appear generous on paper yet hide vague or shifting limits in practice. With compute-based usage limits, Google gains a clearer framework to sell higher capacity at each step, while customers get a more transparent sense of what each complex query or long-running task truly costs.
Gemini Spark Beta: Testing Always-On AI Agents
The pricing reset arrives alongside Gemini Spark beta, a 24/7 personal AI agent inside the Gemini app. Initially reaching trusted testers around mid-May, Gemini Spark will roll out as a beta to U.S. Google AI Ultra subscribers starting the week of May 26. Spark is designed for longer-running, multi-step tasks such as monitoring inboxes, assembling documents, and managing ongoing workflows rather than responding to isolated prompts. Crucially, Google is pairing this autonomy with visible guardrails: Spark requires explicit approval for higher-risk actions, including sending emails, keeping humans firmly in control of sensitive interactions. By anchoring Spark to the Ultra tiers, Google is effectively using its agent capabilities to justify higher capacity plans, testing whether customers will pay more when they can trust an AI agent to operate continuously within clear, predictable boundaries.
How the New Tiers Stack Up Against AI Rivals
Google’s reworked Gemini pricing enters a crowded premium AI bracket where OpenAI and Anthropic already target heavy individual users with high-end plans, and Microsoft’s Copilot Pro undercuts them at USD 20 (approx. RM92) per month. Google AI Pro remains at USD 19.99 (approx. RM92), preserving a wide jump up to the new AI Ultra tier and the still-priciest upper Ultra plan. The question for buyers is whether Google’s higher prices translate into enough extra capacity, integrated tools, and agent-style features like Gemini Spark to justify the upgrade. For power users who routinely hit Pro-level limits, the AI Ultra tier cost may look attractive as a stepping stone. Others may continue to treat Google’s lineup as part of a broader comparison exercise, weighing raw capacity and agent reliability against more affordable rivals and existing workflows.
Who Actually Benefits from Google’s Reset?
The biggest winners from Google’s new structure are users whose workloads routinely strain mainstream plans but do not always justify the highest tier. Developers, technical leads, knowledge workers, and advanced creators now have a more accessible on-ramp to premium capacity, with the AI Ultra tier offering substantial headroom over Pro without jumping straight to the top ceiling. Compute-based limits further help these buyers understand precisely how much value they extract from each plan, especially for intensive coding, data analysis, and media generation. Meanwhile, the upper Ultra plan still targets the heaviest scenarios: extended agent sessions, continuous monitoring, and complex, multi-stage tasks. If the balance between these tiers holds, Google secures a broader premium funnel. If not, price-sensitive users may favor lower-cost competitors or stay on Pro, treating Ultra as a niche option for only the most demanding workflows.