Google is rolling back some of Gemini's restrictive usage limits following widespread user complaints. The company announced it will cap how much quota a single prompt can consume, addressing concerns that users were exhausting their allowance with just one request. Gemini lead Josh Woodward confirmed the changes on social media, acknowledging that the new system needed refinement.

The backlash began after Google I/O 2026 introduced a compute-based usage system replacing fixed message limits. Under this new approach, usage caps apply within rolling five-hour windows plus weekly limits. Users reported hitting these restrictions unexpectedly, with some claiming a single prompt consumed significant portions of their quota or that limits triggered during video generation tasks.

Google's adjustments include exempting failed requests from quota calculations—only successfully completed prompts will count toward limits. The company is also adding detailed usage breakdowns and notifications to help users track consumption. Additionally, Gemini will now remember preferred models across sessions unless manually changed or forced to fallback due to hitting caps.

These changes show Google responding quickly to user feedback, though the underlying compute-based system remains in place. The adjustments should provide more predictable usage patterns for heavy Gemini users while maintaining resource controls.