Gemini App Auto Models and Thinking Levels for Free Users

Google just changed the Gemini app for free users. Starting October 9, 2026, accounts without a paid Google AI plan no longer pick specific models like Flash or Pro. The model picker shows only Auto with Gemini 3 models. At the same time, everyone gets three new thinking or effort levels: Low, Medium, and High.

These updates affect how you get answers on Android, iOS, and the web version of Gemini Apps. Free users trade manual model choice for automatic routing and finer control over reasoning depth.

What Changed for Free Gemini Users

According to Google’s official support page, Auto becomes the default for users without a Google AI plan. Gemini uses auto model selection to send each prompt to the model that suits it best. Most prompts go to Flash-Lite, Google’s fastest model. Prompts that need deeper reasoning may go to Flash or Pro.

If you turn off smart model selection, every response for free users uses Flash-Lite. You cannot manually select Flash or Pro anymore on the free plan. Paid plans keep more choices. AI Plus users keep Flash-Lite and Flash. AI Pro and Ultra retain all three models, and AI Pro gains Deep Think.

The change started rolling out on October 9 for free accounts and reached most users quickly.

How Auto Model Selection Works

Auto decides the model per prompt. Everyday questions about weather, simple summaries, or quick facts stay on Flash-Lite for speed. Complex coding, multi-step math, or research-style prompts can move to stronger models when the system detects the need.

You do not see which model handled each reply. The system handles routing behind the scenes. Turning smart selection off locks you to Flash-Lite only, so leave it on if you want occasional access to Flash or Pro.

This setup matches how many free-tier AI services manage compute. It prioritizes speed for most queries while preserving capacity for harder ones.

New Thinking Levels: Low, Medium, and High

Google replaced the previous Extended option with three effort levels available for each model. These match the thinking levels already in Google AI Studio and Antigravity.

• Low: Quick and efficient responses.
• Medium: Balanced depth for most tasks.
• High: Extra thorough reasoning for complex work.

Higher effort improves the model’s ability to complete tasks and deliver more thorough answers. It also consumes more of your usage limit. Choose Low for routine chats, Medium as the everyday default, and High when accuracy on hard problems matters more than speed.

These levels roll out to free users and all subscribers. Paid users still have access to Deep Think on supported plans for maximum parallel reasoning.

Step-by-Step: Adjust Settings in the Gemini App

Follow these steps on mobile or web. Update the app first for the best experience.

1. Open the Gemini app or go to gemini.google.com and sign in.
2. Start a new chat or open an existing one.
3. Look near the prompt box or model selector for the Auto option and effort controls.
4. Confirm Auto is selected. Free users cannot switch away from it.
5. Select Low, Medium, or High under the effort or thinking level.
6. Type your prompt and send. Note response speed and detail.
7. Check Settings if smart model selection appears as a toggle and leave it enabled unless you want only Flash-Lite.

Test the same prompt at different effort levels to see the difference. High takes longer but often produces clearer step-by-step reasoning on difficult questions.

If you connect other Google apps, review how those extensions behave under the new limits. See our guide on connecting apps to Gemini for setup details: /blog/connect-apps-to-gemini/.

Tips for Free Users

Keep smart model selection on so Auto can route harder prompts upward. Use Low for quick checks and High only when the task justifies the extra limit usage. Break complex jobs into smaller prompts if High burns through your allowance too fast.

Monitor your usage in the app. Free plans have standard limits that reset regularly. If you hit limits often, consider AI Plus for more Flash access or higher overall capacity.

Write clear prompts. Specific instructions help Auto choose the right model and reduce the need for High effort. Include context, desired format, and examples when possible.

For image or file tasks, confirm the feature still works under Flash-Lite defaults. Some advanced capabilities may feel slower or less capable until Auto routes them higher.

What This Means for Different Plans

Free users lose direct model choice but gain automatic optimization and the three effort levels. Most daily use stays fast. Occasional deeper work can still reach stronger models if the system allows it.

AI Plus subscribers face limits on Pro but keep Flash. AI Pro and Ultra see no loss of models and gain or retain Deep Think. The support document outlines exact availability after the October changes.

Check your plan status in the Gemini app settings if the update has not appeared yet. Rollouts can vary by account and region.

Conclusion

The shift to Auto for free Gemini users simplifies the interface while adding practical control through Low, Medium, and High effort levels. Most prompts stay fast on Flash-Lite. Harder ones can still escalate when needed, provided smart selection stays on.

Test the new controls on your next few chats. Adjust effort based on the task and watch your limits. For deeper Gemini features such as connected apps or skills, explore related guides on this site.

Official details live on Google’s support page for Gemini model access changes. Update the app and confirm the settings match your account.

Sources

• Google Support: Changes to Gemini model access and limits. https://support.google.com/gemini/answer/17004136
• 9to5Google reporting on the Auto rollout and effort levels.