Starting October 9, Google's Gemini AI will offer different levels of access based on whether users have a subscription. Those without a subscription will only be able to use Flash-Lite, the basic model, while Flash and Pro will be available to users with AI Plus and AI Pro subscriptions, respectively. The more advanced Deep Think mode, which allows for complex reasoning, will remain exclusive to AI Pro and AI Ultra subscribers. Since May, Google has been measuring Gemini's usage based on computational power, with limits that renew every five hours, up to a weekly cap. This change was designed to manage the system's resources more efficiently. The new changes in October focus more on the type of access rather than the amount. A chart on Google's help page outlines these changes clearly, showing that users without a subscription will only have access to Flash-Lite, while higher-tier subscribers will have access to more powerful models. The changes will start applying on October 9 for users with a personal account but no AI subscription. According to Google's chart, these users will no longer have access to anything other than Flash-Lite. Flash and Pro will be unavailable to them. AI Plus subscribers will keep access to Flash-Lite and Flash but not Pro. Google will send an email to these users specifying the exact date the change affects them. Meanwhile, users with the more expensive AI Pro and AI Ultra subscriptions will retain access to all three models. Google has not provided a detailed explanation for these changes, but company leaders have mentioned since February that computational capacity is a limiting factor. The structure of these limits is similar to those of other major AI providers, such as Anthropic and OpenAI. For example, OpenAI's free version of ChatGPT includes its lightest model, GPT-5.6 Luna, while more powerful features are reserved for paid users. At both Google and OpenAI, the free version offers the basic model, and enhanced capabilities are available through paid subscriptions.