From One Ultra Ceiling to a Tiered Gemini Pricing Ladder
Google’s latest restructuring of Google Gemini pricing introduces a clearer ladder for premium AI access. The company unveiled a new AI Ultra plan priced at USD 100 (approx. RM460) per month and lowered its former top Ultra subscription from USD 250 (approx. RM1,150) to USD 200 (approx. RM920) per month. This creates two distinct Ultra lanes above the existing Google AI Pro plan at USD 19.99 (approx. RM92) per month. The entry-level Gemini Ultra tier offers a cheaper on-ramp to Google’s highest-end consumer bundle, while the upper Ultra option remains the destination for the most demanding workloads. Rather than simply stacking more features at higher AI subscription tiers, Google is now differentiating primarily on capacity and workload intensity, making the premium AI pricing story less about unlocking tools and more about how much serious work users can push through Gemini.
Capacity Over Features: How Compute-Based Limits Change Usage
Alongside the new Gemini Ultra tier, Google is overhauling how access is measured. Instead of daily prompt caps, paid users now face compute-based limits that refresh every five hours until a weekly ceiling is reached, with paid top-up credits planned for heavy workloads. This approach moves Google Gemini pricing closer to a metered infrastructure model: users must think in terms of coding sessions, multi-step research, and media-heavy tasks drawing down a processing budget. Earlier paid plans often looked generous until vague hidden limits surfaced in real-world use. By selling capacity explicitly, Google can segment power users more precisely and monetize sustained, automation-like workloads. For buyers, the trade-off shifts from “Do I get this feature?” to “How much reliable compute do I need for my typical week of AI-assisted work?”
What the New $100 Ultra Tier Actually Buys
The new Gemini Ultra tier at USD 100 (approx. RM460) is positioned as an entry lane for developers, technical leads, knowledge workers, and advanced creators who have outgrown Pro but do not yet need the top Ultra ceiling. This tier brings 20 TB of storage and YouTube Premium, plus allowances in the Gemini app and Google Antigravity that are five times higher than AI Pro. The upper Ultra tier keeps the same feature set but raises usage limits to 20 times Pro, making headroom for longer and more demanding workloads the primary reason to upgrade. In practice, users choosing between AI subscription tiers are now deciding how much sustained throughput they require for coding, complex content creation, and multi-tool workflows, rather than paying more simply to unlock a different set of capabilities.
Gemini Spark Beta: Testing Agents, Guardrails, and Willingness to Pay
Google is using the new pricing structure to launch and test Gemini Spark, a 24/7 personal AI agent inside the Gemini app. Spark is reaching trusted testers and then rolling out as a beta to U.S. Google AI Ultra subscribers, tying agent-style functionality directly to the higher-capacity tiers. Spark is designed for longer-running tasks—monitoring inboxes, assembling documents, and staying active across workflows—where higher usage ceilings become crucial. At the same time, Google emphasizes guardrails: Spark requires explicit approval for high-risk actions such as sending emails, making safety a core part of the value proposition. The experiment is strategic: Google wants to see whether users will pay for longer-running, semi-autonomous agents when boundaries and controls are clearly defined, and whether the new entry Ultra tier can sit comfortably between Pro and the upper Ultra lane.
Competitive Pressures and the Future of Premium AI Pricing
Google’s reset lands in a crowded premium AI bracket. OpenAI already sells high-end plans aimed at heavy users, while Anthropic prices its Claude Max 5x and Max 20x offerings at similar high-end price points. Microsoft’s Copilot Pro at USD 20 (approx. RM92) per month remains a much cheaper baseline, putting pressure on Google to justify the cost of its Gemini Ultra tier options through higher capacity and richer tooling. By formalizing compute-based limits and adding agent-focused features like Gemini Spark, Google is betting that serious users will pay for predictable throughput and robust guardrails. The real test will be how users navigate the expanded ladder: whether they see the new USD 100 (approx. RM460) Ultra entry as a sweet spot for advanced work, or whether they treat Google’s lineup as just another comparison matrix against rival AI subscription tiers.