What you need to know
In response to user feedback, Google is implementing significant adjustments to its Gemini platform, particularly regarding usage limits. Following a wave of complaints from users who found themselves reaching quota limits unexpectedly, the tech giant is rolling out changes aimed at enhancing user experience.
- Google will now impose a cap on how much quota a single Gemini prompt can consume, ensuring that users can maximize their usage of the Pro model.
- Requests that fail will no longer count against users’ quotas, allowing for a more forgiving experience when errors occur.
- To provide greater transparency, Google plans to introduce detailed usage breakdowns, enabling users to track their quota consumption more effectively.
The recent changes come after the announcement at Google I/O 2026, where the company shifted from fixed message limits to a more flexible compute-based usage system. This new framework includes caps within rolling five-hour windows, alongside an overarching weekly limit. However, many users quickly voiced their frustrations on platforms like Reddit and X, reporting that they were exhausting their quotas far too quickly. One user noted that a single prompt consumed a substantial portion of their allowance, while another encountered limits mid-request when asking Gemini Omni to generate a video.
Josh Woodward, a representative from Google, confirmed that the company is taking these concerns seriously. He stated that the new cap on quota consumption per prompt is designed to allow users to derive more value from the Pro model. Additionally, he clarified that only successfully completed requests will count toward usage limits, meaning that failed attempts will not deplete users’ allowances.
Woodward also acknowledged the increased resource demands of more complex tasks, such as Deep Research. To address this, Google will provide more detailed notifications and usage breakdowns, helping users understand how their quotas are being utilized.
In a further enhancement, Gemini will now remember users’ preferred models across sessions. Once a model is selected, Gemini will continue to use it until the user decides to switch or reaches a usage cap that necessitates a fallback to a lighter model.
