Gemini free tier shifts to automatic model picks, adds reasoning effort controls
Google is moving free-plan Gemini app users to automatic model selection. The previously announced change is now being deployed. All users are also gaining controls that adjust how much reasoning effort the model applies.
Google is now rolling out a previously announced adjustment for people using the free version of its Gemini app: instead of selecting a model themselves, they will be routed through an automated choice. The same update gives every user new settings that determine how much effort the model devotes to reasoning. This fits a broader pattern in AI assistants, where providers manage model routing behind the scenes while offering limited tuning options to users.
The shift may affect free-tier users most directly, since their model behavior could change without a manual pick. All users may gain more say over how much computation is spent on reasoning, which could alter response speed, depth, or consistency. Over time, such changes could shape expectations about transparency and control in everyday AI tools, especially for people who rely on free access.