MOUNTAIN VIEW, CALIFORNIA / RankWire.AI / – Google will restrict the selection of Gemini models available to personal accounts without paid subscriptions starting October 9, 2026. Users who do not hold a Google AI plan will still have access to Flash-Lite through Gemini Apps, but access to Flash or Pro will be phased out under the new structure. Google outlined these updates in revised support documentation for its consumer AI services. The changes specifically impact personal Google accounts, not those managed through workplaces or educational institutions.
Customers subscribed to Google AI Plus will retain access to Flash-Lite and Flash after the modifications are implemented on their accounts. However, their subscription will no longer include Pro. Google has not provided a specific date for when all AI Plus users will experience this transition. The company states that affected users will receive an email indicating when their access will change. Meanwhile, AI Pro and AI Ultra subscribers will continue to have access to Flash-Lite, Flash, and Pro.
Flash-Lite is optimized as a quick-response Gemini model suitable for everyday tasks and fast replies. Flash offers additional functionalities with still quick response times across various request types. Pro is designed for more intensive workloads, such as complex mathematics, coding, and advanced analysis, capable of processing text, images, files, and videos while utilizing deeper reasoning. Responses generated by Pro may require more processing time compared to those from other Gemini models.
Wider model access maintained through paid Gemini subscriptions
Google is also rolling out effort settings categorized as low, medium, and high, which can be selected based on user plan. Higher effort levels enable models to allocate more resources for producing detailed answers but also consume more of the account’s usage quota. Subscribers to AI Pro and AI Ultra will gain access to Deep Think, a feature designed for demanding reasoning tasks, which applies maximum parallel reasoning to handle complex requests, according to Google.
These new restrictions follow earlier adjustments to how Gemini Apps calculate usage limits. On May 17, 2026, Google introduced compute-based limits, with allowances resetting every five hours until weekly caps are reached. The calculation takes into account factors such as prompt complexity, features used, and conversation length. Premium features and advanced models tend to deplete these allowances faster than basic requests. The same usage-limit rules were later extended to users under 18 on July 24.
Subscription plans for Google AI vary in their usage allowances
Users without a Google AI subscription are subject to the service’s standard usage limits. AI Plus members receive twice the standard quota, while AI Pro subscribers enjoy four times that amount. AI Ultra offers either five times or 20 times the AI Pro allowance, depending on the subscription. Tools such as Deep Research, media generation, Pro, extended thinking, and Deep Think can utilize larger portions of these limits. Consequently, paid plans not only expand model access but also provide higher usage allowances.
The October updates apply solely to Gemini Apps accessed through personal Google accounts and do not influence other Google AI services. Work and school accounts follow separate access rules, and developer services have their own policies. Users can modify, upgrade, or cancel their Google AI plans via Gemini Apps and Google One. Starting October 9, free personal users will rely on Flash-Lite instead of Flash or Pro, with Google also recommending keeping the Gemini mobile app updated as the new access structure takes effect.