Google has changed how Gemini usage limits are counted after users reported that some requests were consuming quotas too quickly. Reports from The Decoder, 9to5Google, Android Central, and heise online say the company fixed an Omni video bug, stopped charging quota for failed requests, capped usage for complex Gemin...
Google has adjusted Gemini usage limits after users complained that some AI requests were consuming quotas faster than expected.
According to reports from The Decoder, 9to5Google, Android Central, and heise online, the changes follow frustration over Google’s newer compute-based limits for Gemini. Those limits are designed to account for how resource-intensive different requests are, but some users found that their available quota could disappear after a small number of prompts or video generations.
The Decoder reported that a bug in Google’s Gemini app caused one or two Omni video generations to use up an entire quota for some users. 9to5Google also described the issue as an Omni video bug that drained quotas unexpectedly.
Google has fixed that bug, according to the same reports. The Decoder said Ultra members now receive twice as many video generations, while heise online reported that Google is making Gemini quotas last longer and more predictable.
The sources do not provide a full technical explanation of the Omni bug, but they agree that the problem made quota consumption appear much harsher than intended for affected users.
Google is also changing how failed requests are handled. 9to5Google reported that failed Gemini requests will no longer be charged against a user’s quota. Android Central similarly said Google confirmed that failed requests would not consume quota after complaints from users.
That change addresses a basic usability issue: if a request does not complete successfully, users should not lose part of a limited allowance. The reports frame this as one of several immediate adjustments Google is making in response to feedback.
The updates also affect Gemini 3.1 Pro usage. 9to5Google reported that Google is capping how much quota a Gemini 3.1 Pro prompt can consume. Android Central said the change is intended to prevent compute-based limits from being exhausted too quickly by individual requests.
Heise online attributed the planned adjustments to Google’s Josh Woodward, reporting that Google wants Gemini quotas to last longer and become more predictable by limiting quota use for complex Gemini 3.1 Pro inputs.
Compute-based usage limits can be difficult for users to anticipate because a short-looking request may still be expensive to process if it requires complex reasoning, long context, or heavy generation. The reported cap is intended to reduce the chance that one unusually demanding prompt consumes a disproportionate share of a user’s allowance.
9to5Google reported another change: Flash-Lite prompts are being made free. The excerpts provided do not specify whether this applies to every Gemini plan or only particular tiers, so that distinction remains unclear from the available sources.
The move suggests Google is separating lower-cost Gemini usage from requests that require more expensive models or tools. It may also help users preserve quota for more demanding tasks, though the sources do not quantify how much quota this will save for typical users.
Google also plans to provide clearer information about Gemini usage. The Decoder reported that Google intends to add more transparency around other usage, while Android Central said more detailed usage breakdowns are planned.
That transparency may be important because compute-based quotas are less intuitive than simple message caps. Under a message cap, users know roughly how many prompts remain. Under a compute-based system, the same number of prompts can consume very different amounts of quota depending on the model and task.
The reports indicate that Google is responding quickly to criticism, but they do not say whether the company will return to simpler limits or keep compute-based accounting as the long-term approach. For now, the confirmed changes reported by The Decoder, 9to5Google, Android Central, and heise online are narrower: fixing the Omni video quota bug, exempting failed requests, capping high-cost Gemini 3.1 Pro prompts, making Flash-Lite prompts free, and promising more detailed usage information.
Google has adjusted Gemini usage limits after users complained that some AI requests were consuming quotas faster than expected.
According to reports from The Decoder, 9to5Google, Android Central, and heise online, the changes follow frustration over Google’s newer compute based limits for Gemini.
Those limits are designed to account for how resource intensive different requests are, but some users found that their available quota could disappear after a small number of prompts or video generations.
Continue reading
