Changes to Gemini model access and limits (support.google.com)

Gemini has announced significant changes to its model access and usage limits, transitioning to a compute-based system that refreshes every five hours, taking into account prompt complexity, feature usage, and chat length. This new model primarily affects users aged 18 and over, with paid subscribers enjoying higher usage thresholds compared to free users. The update introduces various tiers: without a plan, AI Plus, AI Pro, and AI Ultra, which offer progressively higher limits, ensuring that premium features such as media generation, deep research capabilities, and extended thinking tools consume usage more efficiently. This shift is particularly notable for the AI/ML community as it emphasizes the importance of optimizing resource utilization based on computational demands. By aligning usage limits with the complexity of tasks performed, Gemini aims to enhance user experience while managing server loads effectively. These changes signal a growing trend towards personalized consumption models in AI services, which could lead to more tailored interactions and advancements in AI applications across diverse fields.