Google Restricts Advanced Gemini AI Access for Free and Some Paid Users

Google is limiting access to its advanced Gemini AI models for both free users and some paid subscribers. Starting October 9, free users will only be able to access the basic Flash-Lite model, with access to the more powerful Flash and Pro models being discontinued for them. This means free users will lose the ability to use Gemini 3.6 Flash and Gemini 3.1 Pro, being left with Gemini 3.5 Flash-Lite, Google's lightest and most economical model designed for quick, everyday tasks. The Pro model is intended for more complex tasks like coding, problem-solving, and analyzing text, files, and images.

The changes will also affect subscribers of the mid-tier Google AI Plus plan (priced at 19 shekels per month or 190 shekels annually). These users will retain access to Flash-Lite and Flash but will no longer be able to select the Pro model. The exact date for this change for AI Plus subscribers has not been announced, but they will receive an email notification. Subscribers of the AI Pro (74.9 shekels per month or 749.9 shekels annually) and AI Ultra (390 shekels per month) plans will continue to have access to all three model tiers.

Additionally, the AI Pro plan will gain the "Deep Think" mode, previously exclusive to Ultra subscribers. This mode allows the model to spend more time and computational resources on generating responses, exploring multiple solutions simultaneously. Google also plans to allow users to select one of three computational cost levels, low, medium, or high, for each model. Higher levels yield more detailed results but consume more of the user's quota. Similar settings are available from competitors like Anthropic's Claude and OpenAI's ChatGPT.

Last week, Google unveiled Gemini 4 Argon, its next-generation model. Currently, access is restricted to a select group of experts, cybersecurity organizations, and the U.S. administration, which Google attributes to the model's potent cybersecurity capabilities.