Google Doubles Gemini Omni Usage Limits for AI Ultra Subscribers Following a Quota Bug
In Focus
- The video generation bug exhausted usage quotas fast
- Google has doubled video generation quotas for users to fix the usage issue
- The tech company introduced compute-based limits during the I/O 2026 event
Google has resolved the Gemini Omni quota bug that had triggered high credit consumption after facing user backlash. The video generation bug exhausted usage quotas after one or two requests, leading to unusually high consumption. Google also informed users that unsuccessful AI requests will not be charged against user quotas and that usage will be based on completed actions.
What Triggered the Gemini Quota Issue?
Google announced plans to shift the Gemini app from daily quota limits to compute-based limits during the Google I/O 2026 event. This shift has caused weekly usage quotas to run out faster due to high computational data power required to complete video generation requests. Several users raised concerns on social media.
Some users complained that a single attempt to create an AI avatar video using the Gemini Omni AI tool consumed a five-hour usage allowance in minutes and failed to generate the output well. Google noted the complaints and committed to look into the issue.
The company has fixed the Gemini Omni quota issue and introduced several changes for Google AI plan subscribers. Google adjusted usage quota for Gemini Omni users days after it expanded limits for Gemini Pro models following severe backlash over its new compute-based caps.
What Adjustments has Google Announced?
As part of addressing the Gemini AI usage limits problem, the company has doubled video generation quota for Google AI Ultra subscribers. The tech giant has also introduced multiple changes across its paid plans to make usage more predictable for customers.
Google’s VP Josh Woodward announced that the tech giant will no longer charge users for failed requests and that usage quotas will only apply to successful tasks. The VP said that Flash-Lite prompts will be offered free of charge and will not be charged against usage limits.
Concerning Gemini 3.1 Pro prompts, the VP noted that they were depleting usage quotas fast. Google is now capping quotas that a single prompt can consume. This change will allow users to complete more tasks using the Pro model.
Google to Develop New Limits for Heavy Tasks
The company also noted tasks such as Deep Research also require more computational power. To meet this need, Google is developing detailed usage plans and notifications to enable users to manage their limits better.
The Google VP noted that Gemini will now remember the model that a user selects. The company will only change the selected model if the option is adjusted manually or when a usage cap activates resumption to a lighter model.
Google’s quota update is aimed at rebuilding user trust as AI competition grows. With more users relying on generative AI tools, demand for clearer policies, reliable performance, and fair pricing continues to rise
