Google is sharply restricting which Gemini models free users and budget subscribers can access, according to an updated support document first reported by 9to5Google. Starting October 9, users without a paid Google AI subscription will be limited to the Gemini 3.5 Flash-Lite model, losing access to both Gemini 3.6 Flash and Gemini 3.1 Pro in the Gemini app.
The change, which multiple outlets including TechRepublic and PCMag covered this week, is the most significant reshuffle of Gemini's plan structure since Google introduced compute-based usage limits in May. It effectively turns the free tier into a single-model experience and pushes anyone who wants capable general-purpose models toward paid plans. For more context on this story, see our ongoing AI news.
What changes for each plan
The new access rules differ sharply by tier. Free users will only be able to use 3.5 Flash-Lite, the smallest and fastest model in the current lineup. AI Plus subscribers, who pay $4.99 per month, will be limited to Flash-Lite and Flash, with Gemini 3.1 Pro no longer available on the plan. Google said those subscribers will receive an email explaining when the changes take effect for their accounts.
There are no regressions for AI Pro and AI Ultra subscribers. AI Pro, the $19.99 per month plan, retains access to the full current model lineup — and is gaining a significant new capability, as detailed below.
Deep Think moves down-market
As part of the reshuffle, AI Pro subscribers will get access to the Deep Think option, which Google describes as delivering maximum parallel reasoning on complex problems. Deep Think was previously reserved for AI Ultra customers on the $99.99 and $199.99 per month plans, making its arrival on the $19.99 tier a notable sweetener as Google removes Pro-model access from cheaper plans.
The move follows a familiar play book: constrain what entry-level plans can do while adding premium features higher up, nudging users up the pricing ladder. With Flash-Lite as the only free option, users who rely on Gemini for substantial daily work — coding, long document analysis, complex reasoning — will find the free tier considerably less capable from Friday.
Effort levels arrive in the app
Google is also preparing to add new low, medium and high effort levels for each available model in the Gemini app this October, according to 9to5Google. Higher effort settings increase a model's ability to complete tasks and produce more thorough answers, but consume more of the user's usage limit.
The three-level system matches what already exists in Google AI Studio and the Antigravity IDE, and it replaces the current simple toggle for Extended thinking in the consumer app. For Google, the granular system also helps manage compute demand: by letting users choose how much reasoning power each query receives, the company can reserve its most expensive inference capacity for the subscribers who actually need it. For paid subscribers, effort levels add a new dimension of control; for free users, they matter less given the single-model restriction.
Where Gemini 4 Argon fits
The tier reshuffle comes ahead of a broader model transition. Google announced Gemini 4 Argon, its new frontier model, on September 30, and said it would initially roll out to AI Ultra subscribers. Google has not yet specified whether Argon will be classified as a Pro-class model or sit in a new tier above it, leaving some uncertainty about which plans will eventually get access.
That uncertainty has not gone over well with some subscribers. Comments on the 9to5Google report show AI Pro customers questioning what they are paying for if 3.1 Pro disappears from lower tiers and the new flagship remains an Ultra exclusive.
What users should know before October 9
The changes take effect on October 9 for free users, with AI Plus subscribers receiving individual email notices about their transition timing. Users who depend on Flash or Pro within the app will need to upgrade to AI Pro at $19.99 per month to keep comparable capability, or adjust to the lighter Flash-Lite experience.
The reshuffle underscores how quickly the economics of consumer AI are shifting. Free access to frontier-adjacent models was a customer acquisition strategy in 2025; in late 2026, Google is betting that enough users will pay for capability that the free tier can be stripped down to its lightest model without slowing subscriber growth.
The timing also looks deliberate. Gemini 4 Argon's initial rollout to AI Ultra, combined with Deep Think's arrival on AI Pro, points to a rebuilt tier structure around the next model generation: entry plans keep fast, lightweight models, the mid tier picks up reasoning features, and the top tier gets the new flagship first. That gives Google a clearer upsell path at every price point — and makes October 9 the date the old Gemini plan ladder effectively disappears.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
