Gemini’s free tier lost its model picker on October 9, but not in the way Google’s own announcement suggested it would. Free users without a Google AI plan now see a single entry in the model list: Auto. Google’s support page and the app’s own text both say Auto can still route a prompt to Flash or Pro. That happens when the request needs deeper reasoning. What free users actually lost is the ability to choose those models by hand.
What we reported earlier, and what changed
On October 4 androidpure published Gemini Free Users Lose Flash and Pro Models on October 9. That post said free accounts would be limited to Flash-Lite, full stop. It was based on Google’s support-page wording as it read at the time.
Google has since expanded that same support page. The page now describes Auto as the default setting for users without a Google AI plan. It also states that turning smart model selection off sends every response on a free account to Flash-Lite. So the earlier framing was not wrong about Flash-Lite doing most of the work. It was too absolute about Flash and Pro being gone entirely.
How Auto actually decides
Google’s in-app text spells out the logic in plain terms:
Gemini will use auto model selection to send each prompt to the Gemini model that suits it best. Most prompts will go to Flash-Lite, our fastest model. Prompts that need deeper reasoning may go to Flash or Pro.
That text is Google’s own product copy, as reported by 9to5Google. Read it closely and the tradeoff is clear. Most prompts will go to Flash-Lite, and the decision about which ones do not is no longer yours. Before this change a free user who wanted Pro on a specific question could just select it. Now the router makes that call for you, against criteria Google describes only in general terms.
That is the part worth being skeptical about. Auto is presented as a convenience feature that saves you from thinking about model names. It also removes a control free users had last week.
Effort levels replace the Extended toggle
The same update swaps the old Extended thinking toggle for three effort levels: low, medium and high. Google documents this on the same support page, describing it as a per-model setting. Unlike the model-picker change, effort levels are going to all subscribers rather than just the free tier. The naming matches the thinking levels terminology Google already uses in AI Studio and Antigravity.
The rest of it, in brief
- Rollout reach: 9to5Google reported the change had reached the vast majority of its own free accounts by the evening of October 9.
- Google AI Plus tier: 9to5Google says Plus subscribers will soon be capped at Flash-Lite and what it calls 3.8 Flash. Google’s page ties that timing to individual email notices, so do not read October 9 as the Plus date.
- Inconsistent labels: the in-app picker currently describes the free option as using Gemini 3 models. Google’s own documentation referred to the same thing as 3.5 Flash-Lite, per 9to5Google. The mismatch is Google’s, not a bug on your phone.
- Who is covered: Google’s page refers to personal Gemini accounts and states no regional limit. That is not the same as Google confirming a global rollout.
If you are on a free account and have not seen the change yet, it is arriving account by account rather than all at once. The practical check is simple: open the model list in the Gemini app and see whether anything other than Auto is still there. If Auto is all you get, you can still pin every reply to Flash-Lite by switching smart model selection off.
Sources: Google, 9to5Google
