Gemini compute-based usage limits frustrate users

Google has changed how it meters usage inside the Gemini app, and some paying subscribers say the new system makes heavy use of video and other premium features run out faster than expected. The change took effect on May 17, according to Google’s Gemini Help documentation, which now says limits are based on compute use rather than a simpler count of prompts or chats. Gadgets 360 reported on May 22 that users began complaining on social media after hitting limits sooner than before, particularly when using media-generation tools. Google’s public documentation says the limits depend on prompt complexity, the features used and the length of a chat. ### What exactly changed in Gemini’s usage system? Google’s help page says Gemini Apps now use “compute-based usage limits” that determine how much a person can interact with tools and features inside the service. The company says those limits refresh every five hours until a user reaches a weekly limit, and it warns that limits “may change” and can also be affected by testing or availability. (support.google.com) May 17, 2026 is the date Google attached to the change notice in its help documentation. Gadgets 360 reported that Google informed users by email rather than through a public blog post or social-media announcement, and said the email described limits that factor in prompt complexity, features used and chat length. ### Why are subscribers focusing on video generation? (support.google.com) Gadgets 360 said the sharpest complaints came from users trying premium features such as video generation and deep research. In its own testing, the publication said one Veo 3 video consumed 26% of a five-hour usage window on a Gemini AI Pro account. The same report said one Nano Banana Pro image used 1% of the five-hour allotment, while one deep research task using Gemini 3.1 Pro (Extended) used 5%. (support.google.com) Google has been pushing more video creation features since I/O 2026. On May 19, Google said Gemini Omni was being released across its products and described it as a model that can “create anything from any input, starting with video,” while also highlighting Google Flow for creative work. ### What do paying users actually get with AI Pro? (gadgets360.com) Google One’s pricing page lists Google AI Pro at $19.99 a month in the United States. The company says that tier includes 5 TB of storage, access to Gemini’s Pro model, Deep Research and “4x higher” Gemini app usage limits than accounts without a Google AI plan. Google’s support page does not publish a standard unit showing how many prompts, minutes or videos each tier guarantees. (blog.google) Instead, it describes relative access levels: standard limits for users without an AI plan, 2x for AI Plus, 4x for AI Pro, and 5x or 20x higher than AI Pro for AI Ultra depending on subscription. ### Why are users saying the limits feel tighter now? (one.google.com) Gadgets 360 reported that users were frustrated because the company did not provide a standardized measure of compute consumption for each response or media-generation task. That leaves subscribers trying to infer cost from a dashboard after each use, rather than knowing in advance what a Veo 3 clip, image generation request or deep research run will consume. (support.google.com) The publication also reported that Google is extending a similar product-based usage model to Flow and Antigravity. That matters because Google has been broadening access to creative and agentic tools across its consumer AI products at the same time it is tightening how usage is counted. ### Where can users check the rules now? (gadgets360.com) Google directs personal-account users to its Gemini Help page for the current rules on usage limits, age requirements and plan availability. The Google One plans page separately lists pricing and the relative access levels for AI Plus, AI Pro and AI Ultra. As of May 22, those pages are the main public references for how the company describes the new system. (support.google.com) (gadgets360.com)