Google is pulling its more capable models out of the free tier of the Gemini app. From October 9, users without a paid Google AI plan will be limited to Flash-Lite, the company’s smallest and fastest model, while Flash and Pro move behind a subscription.
The change is being flagged to users through an in-app popup, which says Flash and Pro will soon move to paid plans, while unlimited access to the latest Flash-Lite will continue. Google’s support page says the changes apply to Gemini Apps on personal accounts and begin in October 2026.
Who Gets What
According to Google, model availability after the change will look like this:
| Google AI plan | Flash-Lite | Flash | Pro |
|---|---|---|---|
| No plan (free) | Yes | No | No |
| AI Plus | Yes | Yes | No |
| AI Pro | Yes | Yes | Yes |
| AI Ultra | Yes | Yes | Yes |
The changes take effect on October 9 for users without an AI subscription. AI Plus subscribers will get an email explaining when the changes will reach them, which suggests their rollout may follow a different schedule.
Notably, AI Plus users won’t get Pro access under the new setup, which makes AI Pro the cheapest plan that includes all three model tiers.
Within each available model, users can also pick an effort level of low, medium or high. Higher effort improves the model’s ability to complete tasks and produces more thorough answers, but eats into the user’s usage limit faster. AI Pro and AI Ultra subscribers also get the Deep Think option, which Google describes as maximum parallel reasoning.
Usage Limits Were Already Changing
The model restrictions follow a bigger overhaul of how Gemini meters usage. Since May 17, Gemini has run on compute-based limits rather than fixed daily prompt counts, a shift Google announced around the time it launched Gemini 3.5 Flash at I/O. That change reached users under 18 on July 24.
Under the compute-based system, limits refresh every five hours until a user hits a weekly cap. Usage is calculated from the complexity of the prompt, the features used and the length of the chat. Paid plans get multiples of the standard limit: AI Plus gets 2x, AI Pro gets 4x, and AI Ultra gets 5x or 20x the Pro limit depending on the subscription.
Google also notes that premium models and features burn through allowance faster. That includes image, video and music generation, Deep Research, the Pro model, and extended thinking or Deep Think. Moving Flash and Pro off the free tier means free users will no longer have access to two of the models that consume the most compute.
Flash Has Been Google’s Headline Act
The timing is notable because Flash is where much of Google’s momentum has been this year. Google released Gemini 3.8 Flash on September 2, the latest in a rapid run of Flash updates that began with Gemini 3.6 Flash and Gemini 3.5 Flash-Lite in July. In between, Gemini 3.7 Flash was described by Sundar Pichai as Google’s fastest-growing model ever.
The Pro tier, meanwhile, has been quieter. Gemini 3.5 Pro, announced at I/O, never shipped, and reports suggest it has been shelved in favour of skipping ahead to Gemini 4. Some X users claim a Gemini 4 Pro is already being tested in Arena, though Google hasn’t confirmed anything. Earlier this week, Google introduced Gemini 4 Argon to partners, but hasn’t made it available to the general public.
For free users, the practical effect is that the models Google has been promoting most actively will soon sit behind a paywall, leaving Flash-Lite as the only option. Anyone who wants Flash or Pro will need to subscribe to a Google AI plan, which Google says can be started, upgraded, changed or cancelled at any time from within the Gemini app.