Google Is Restricting Features in Gemini’s Free Version

Google has announced new plan restrictions for users of its Gemini artificial intelligence service. As a result of the changes, the free version and the AI Plus subscription will lose access to some artificial intelligence models. Users of the more expensive AI Pro and AI Ultra plans, by contrast, will gain additional capabilities.

Starting October 9, 2026, users without a paid subscription will only be able to use the Gemini 3.5 Flash-Lite model. Access to the Gemini 3.6 Flash and Gemini 3.1 Pro models will be unavailable to free users.

The changes will also apply to the AI Plus plan, which costs $4.99 per month. These subscribers will only be able to use the Flash-Lite and Flash models, while the Pro model will be unavailable to them. Google will notify AI Plus users by email of the exact date when the changes will take effect.

The list of models available to AI Pro and AI Ultra subscribers will not change. However, AI Pro users will gain access to Deep Think mode, which is currently available only with the more expensive AI Ultra plan. This mode allows the model to use more computing resources to solve more complex tasks.

Google also plans to divide the depth of reasoning Gemini uses when generating responses into three levels in October: low, medium, and high. When a higher level is selected, the model will be able to devote more computing resources to the response. At the same time, the user’s existing limits will be used up more quickly.

As for the Gemini 4 Argon model, Google said that AI Ultra subscribers will be given access to it initially. The company has not yet disclosed whether the new model will replace the existing Pro model or be offered as a separate option in Gemini.