Google to Limit Free Gemini Users to Flash-Lite Starting Oct. 9, ZDNet Reports
Google to limit free Gemini users to Flash-Lite on Oct. 9, ZDNet reports; AI Plus subscribers lose Pro, AI Pro gains Deep Think.
The change also affects Google's cheapest paid tier. ZDNet reported that subscribers to the $5-per-month AI Plus plan will lose access to the Pro model on the same date. They will be left with the Flash and Flash-Lite variants. A free user who needs Flash would have to pay for AI Plus, while an AI Plus user who needs Pro would have to move to a Pro plan, according to the report.
The $20-per-month AI Pro plan will gain access to the Deep Think reasoning mode. ZDNet reported that Deep Think is currently reserved for AI Ultra users, who pay $100 or $200 per month. The change gives AI Pro subscribers a feature that had been limited to the most expensive plans.
Gemini users can currently access the service through free access, AI Plus, AI Pro and two AI Ultra plans. ZDNet reported that each plan currently offers three models: 3.8 Flash, 3.5 Flash-Lite and 3.1 Pro. Flash is described as a generalist that can handle common tasks such as providing information, summarizing files, generating and analyzing images, and chatting. Flash-Lite is faster but less capable, suited to repetitive tasks but not deeper knowledge and reasoning. Pro is the most analytical model, used for difficult problems in areas such as math, logic and coding.
Under the current setup, Flash is typically the default and go-to model for the least expensive paid plans, while free users typically see Flash-Lite as the default because their access to the other two models is limited, according to ZDNet. The report said Flash-Lite should suffice for most general tasks but not for deep reasoning or complex problem solving.
Google also says each model will allow users to choose an error level among low, medium and high, ZDNet reported. Higher levels can complete tasks and provide more thorough responses, but they consume more of a user's limits. Google uses a compute-based limit rather than a specific number of requests or conversations, with limits refreshing every five hours until a weekly quota is reached. The limits depend on prompt complexity, features used and chat length, according to the report.
Paid users have higher limits than free users, and premium models and tasks use limits faster. ZDNet reported that generating images or videos, using Deep Research or Deep Think, and tapping the Pro model will consume limits more quickly than everyday chats and requests. The report added that Google AI subscriptions are billed monthly, so users can switch between free access and paid plans month by month if their current level is too limited.