Google narrows Gemini app model access: users without a plan keep only Flash-Lite
Google is changing which models personal-account users can reach in the Gemini app starting in October 2026. For people without an AI subscription the switch begins on October 9, after which they can use Flash-Lite alone. Usage limits run on a compute-based system that refreshes every 5 hours up to a weekly cap.
A help center page lays out the new model table
Google announced the change on the "Changes to Gemini model access and limits" page of its Gemini Apps help center. The page says that from October 2026 the set of models available on each plan shifts for anyone using Gemini with a personal account. The timing depends on the plan. Users with no Google AI subscription see the new arrangement start on October 9, while AI Plus subscribers are told to expect an email setting out when it applies to them; the page gives no date for that group.
Android Headlines reported the change on October 3, writing that free users lose Flash and Pro from October 9. Webrazzi covered it for Turkish readers on October 5. The Google help page carries no visible publication date, so we treat the October 3 Android Headlines report as the news date.
Four plans, three models and a limit that resets every 5 hours
The availability table on the page sets out the post-change picture. Without a plan, users get Flash-Lite only. AI Plus covers Flash-Lite and Flash. AI Pro and AI Ultra cover all three models: Flash-Lite, Flash and Pro. Subscribers on AI Pro and AI Ultra can also pick Deep Think, which Google describes as an option for maximum parallel reasoning. On every model a user has access to, an effort setting of low, medium or high is available. Google says the higher settings help the model finish tasks and give fuller answers, at the cost of eating into the limit faster.
The page also sets out how limits work. Webrazzi notes that this compute-cost-based quota system was introduced in May and that the October changes follow it. Google describes "compute-based usage limits that will refresh every 5 hours until you reach your weekly limit". Usage is weighed by how complex the prompt is, which features are involved and how long the chat runs. The tiers are expressed as multipliers: no plan sits at the standard limit, AI Plus at twice that, AI Pro at four times, and AI Ultra at either 5 or 20 times the AI Pro level depending on the subscription. Premium work that drains the limit faster includes image, video and music generation, Deep Research, the Pro model, and extended thinking and Deep Think. The page also notes that an earlier round of changes to limits and model availability took effect on May 17, 2026 for users over 18 and on July 24, 2026 for users under 18.
Small teams running on the free tier are left with the lightest model
Small teams that lean on free Gemini for content drafts, ad copy or quick research will be working with the lightest model after October 9, which can change both output quality and how they plan their workflows. Android Headlines frames it as free users losing Flash and Pro, and Google's own table now places both models on paid plans only. As Google puts it, "Paid users have higher limits than users without a Google AI subscription." The new table extends that gap to the choice of model itself.
A compute-based limit that resets every 5 hours is a planning question of its own. A long conversation, a complex prompt or a feature such as image generation can fill the window sooner, and the weekly cap means a heavy day has to be balanced against the rest of the week. That makes it more important for a business to be clear about which tasks truly need which model tier and which effort level. We covered an earlier step in the Flash line on August 15; that piece is separate from this access change.
The sources say nothing about Türkiye, so this section is commentary
None of the three sources contains anything specific to Türkiye. There is no statement on whether users in the country follow the same dates, no local timetable and no country-level exception. Webrazzi reports the news in Turkish but adds no Türkiye-specific detail. What follows is therefore not a fact drawn from the sources but explicitly UNALSOFT commentary.
Our reading is that, since Google published this as a general plan update, it is reasonable for businesses in Türkiye using Gemini on personal accounts to prepare against the same table. The unknowns deserve the same clarity. The sources do not settle when the change will show up for users in Türkiye, when it takes effect for AI Plus subscribers, what the limits amount to in absolute terms, or whether work accounts are affected. The announcement covers personal accounts, so teams on company accounts should check the information in their own admin consoles.
UNALSOFT's take
We read this less as a subscription notice and more as a reminder of how fragile it is to hang a workflow on a single chat window. If a team built its daily routine around how one model behaves on the free tier, that model leaving the plan changes the work itself. The practical response is plain: write down which task needs which model tier, choose effort levels and limit spend against that list, and keep critical work from depending on one interface's plan changes. That is why in our agentic AI work we treat sorting tasks by model tier as its own step. Because the sources carry no data about Türkiye, this article does not predict a local timeline.
Does your team's AI work depend on one plan staying the same?
A short conversation is enough to sort your tasks by model tier and set up an AI workflow that holds up when plans change.