Google recently announced a series of changes to its AI subscription offerings, most notably a price reduction for its top-tier AI Ultra plan from $200 to $100 per month. However, the company quietly implemented significant modifications to its AI Pro plan that have sparked controversy among users.
The new credit system
Instead of fixed token allowances, Google now employs a compute-based credit system. Each user action consumes a variable number of credits depending on the complexity of the prompt, the features used, and the length of the chat session. This means that a simple text prompt might use a few credits, while a multi-modal request or long conversation could burn through a large portion of the monthly quota.
According to early adopters on Reddit, a single text prompt consumed 13% of their total credits, leading to accusations that the plan is now effectively a scam. Users report that the new limits make the Pro plan barely usable for moderate workloads.
Five-hour windows and weekly caps
Following in the footsteps of Anthropic's Claude, Google now enforces usage windows of five hours. After each window, the credit counter resets, but only until the user reaches the weekly limit. This system aims to prevent spikes in usage but adds complexity for users trying to manage their consumption.
The same limits apply to other Google AI products beyond Gemini, including Antigravity and Flow. It remains unclear whether the context windows—the amount of text the model can remember over a conversation—have been reduced as well.
Background and context
These changes come just days after Google was spotted testing weekly limitations. The company has now confirmed that the test results are being rolled out to all Pro subscribers. Google stated that the new approach is more equitable, allowing the company to allocate resources based on actual computational load rather than crude token counts.
However, the lack of transparency has drawn criticism. Google does not specify the exact number of credits or tokens Pro users receive, only stating that the allowance is four times that of the free tier. This vagueness makes it difficult for users to predict their usage and compare plans.
The announcement was overshadowed by the I/O 2026 conference, where Google focused on the price drop for the Ultra plan. The company introduced a new $100 per month tier between the original $200 Ultra and the $20 Pro, with fewer perks and lower limits. This strategy suggests Google is trying to segment its user base more precisely, but the stealthy nerf of the Pro plan has angered loyal subscribers.
Competitive landscape
The AI industry is seeing a shift toward compute-based pricing. Anthropic's Claude recently adopted a similar model, and OpenAI is rumored to be testing dynamic cost allocation. While Google claims this is fairer because complex requests should cost more, users argue that the practical impact is a reduction in value without compensation.
For power users who rely on the Pro plan for daily work, the new limitations may push them toward the Ultra plan or alternative services. For casual users, the free tier remains unchanged, but the gap between free and paid has widened considerably.
Google is also discontinuing the 1,000 free AI credits for Flow that were previously included with the Pro plan. The company asserts that user experience should not change, but many users beg to differ, as they now have to carefully ration their consumption to avoid hitting caps prematurely.
How to track usage
Users can monitor their credit consumption through the Usage Limit option in Gemini's settings. Google advises checking regularly, as usage limits can change without prior notice.
The broader implications of this credit system are yet to unfold. If widely adopted, it could reshape how AI services are priced and consumed. However, for now, the immediate reaction from the community is frustration. Many feel that Google is using a PR-friendly price cut on the Ultra plan to distract from a hidden downgrade of the Pro plan.
Historically, Google has a pattern of quietly adjusting terms of service and subscription benefits. This move aligns with that trend, but the timing—amidst a major AI push—suggests that the company is struggling to balance profitability and user satisfaction in the competitive AI market.
As the AI landscape evolves, transparency about resource allocation will become increasingly important. Users demand clear metrics and predictable costs. Google's current approach may win in the short term by shifting heavy users to more expensive plans, but it risks alienating the developer and power user communities that are essential for ecosystem growth.
Source: Android Authority News