Google is changing how users access its Gemini AI models, and the upcoming update could make a noticeable difference for people using the free version or the AI Plus subscription.
Starting October 9, 2026, free users are expected to lose access to the Flash and Pro models, leaving Flash-Lite as their available option. AI Plus subscribers will also lose access to Pro, while AI Pro and AI Ultra users will continue to get access to all three model families.
There is some good news for paying subscribers, too. Google is bringing its Deep Think reasoning feature to AI Pro, which could make advanced problem-solving tools available to more users. The company also appears to be preparing additional controls that would let users choose how much reasoning effort Gemini puts into a response.
Here’s what is changing and how it could affect different Gemini plans.
Also read: Google Announces Gemini 4 Argon With Longer Output Limit: See Features, Availability
Google Gemini Free Users Will Lose Flash and Pro Access
According to an updated Google support document, the new model-access restrictions are scheduled to begin on October 9.
Under the reported changes, users who don’t have a paid Google AI subscription will be limited to Gemini Flash-Lite. Access to the regular Flash and Pro models will be removed from the free tier.
Flash-Lite is designed for speed and everyday tasks, making it suitable for activities such as basic questions, summaries, and brainstorming. However, users who rely on more advanced models for complicated requests may notice the difference.
The change doesn’t mean free users will lose Gemini altogether. They will still be able to use the assistant, but their choice of AI models will become more limited.
AI Plus Subscribers Will Also Lose Gemini Pro
Google AI Plus subscribers are facing a separate restriction.
The $4.99-per-month plan will retain access to Flash-Lite and Flash, but Pro will no longer be included under the reported changes. Google is expected to notify affected subscribers by email explaining when the restriction will apply to their accounts.
This could matter most to people who regularly use Pro for demanding tasks, including complex reasoning, coding, and detailed analysis.
Users who mainly rely on Gemini for everyday questions and writing assistance may find Flash sufficient for many of their needs. Those who specifically want Pro access, however, may need to consider a higher subscription tier.
Gemini Model Access: What’s Changing?
Here’s a quick comparison of the reported model availability after October 9.
| Google AI plan | Flash-Lite | Flash | Pro |
|---|---|---|---|
| Free | Available | Removed | Removed |
| AI Plus | Available | Available | Removed |
| AI Pro | Available | Available | Available |
| AI Ultra | Available | Available | Available |
These are the reported model-access changes, rather than a guarantee of unlimited usage. Google can still apply usage limits based on the plan, request complexity, and available capacity.
Google AI Pro Is Getting Deep Think
One of the more interesting developments is the planned expansion of Deep Think to Google AI Pro.
Deep Think is designed for more demanding questions that benefit from extended reasoning. Rather than focusing only on producing a quick answer, it uses parallel reasoning to work through complex problems.
This can be useful for tasks that involve multiple steps, careful analysis, or challenging technical questions.
Until now, Deep Think has been associated with Google AI Ultra. Bringing the feature to AI Pro would give more subscribers an opportunity to use Google’s advanced reasoning capabilities without paying for the highest subscription tier.
Google AI Pro costs $19.99 per month in the US, according to the supplied report. AI Ultra continues to offer Deep Think through its higher-priced subscription options.
The feature requires the Pro model, and its availability and usage limits may depend on the subscriber’s plan.
Gemini May Get Low, Medium, and High Thinking Controls
Google also appears to be working on new controls that could give users more say in how Gemini processes their requests.
The proposed options are Low, Medium, and High thinking effort. If introduced, these settings would allow users to choose how much reasoning effort the model should put into a task.
For example, a straightforward question might not need extensive reasoning. A complicated maths problem or technical analysis, on the other hand, could benefit from a higher setting.
The trade-off is usage. More intensive reasoning can consume a larger portion of a user’s allowance, so selecting the highest setting for every prompt may not be practical.
Gemini already offers Extended Thinking for complex problems. Similar controls are also available in Google AI Studio and Antigravity, making the possible addition to the Gemini app a logical next step.
However, the Low, Medium, and High options should be treated as a feature under preparation until Google confirms their wider availability.
Why Is Google Changing Gemini Model Access?
The changes follow Google’s move toward compute-based usage limits.
Under this approach, Gemini usage isn’t determined only by the number of messages a user sends. The complexity of a prompt, the model selected, the length of a conversation, and the features used can all affect how much of a usage allowance is consumed.
A simple question generally requires less processing than a lengthy request involving advanced reasoning or complex analysis.
Restricting model access by subscription tier gives Google another way to manage these differences. It also creates a clearer distinction between free access and the capabilities available through paid plans.
For users, the practical result is that choosing a subscription may become more important if they regularly depend on a particular model.
Google Gemini 4 Argon Is Also on the Way
Alongside the subscription changes, Google is working toward a wider rollout of Gemini 4 Argon, its newer AI model focused on complex professional tasks, software engineering, and cybersecurity.
Google announced Argon on September 30, 2026, with initial access restricted to trusted cybersecurity professionals through its Fairwind programme. The company says it is gathering feedback and improving safeguards before expanding availability to developers, enterprises, and consumers.
Argon is intended for demanding tasks that require sustained reasoning over multiple steps. Its cybersecurity capabilities are a particular focus, including identifying, validating, and helping fix software vulnerabilities.
The model is expected to reach paid API customers and AI Ultra subscribers as access expands. However, Google has not confirmed whether Argon will eventually be included in AI Pro or offered through a separate subscription tier.
For now, its wider release and final subscription arrangements remain unclear.
What Should Gemini Users Do Now?
If you use Gemini regularly, it is worth checking which model you currently rely on and whether it will remain available under your plan.
Free users who mostly ask everyday questions may find Flash-Lite sufficient for their needs. AI Plus subscribers should consider whether losing Pro access would affect their regular workflow.
AI Pro users, meanwhile, may benefit from the planned Deep Think expansion if they frequently work on complex problems. Still, it is sensible to check the feature’s availability and usage limits before changing subscriptions.
You should also review the current plan details rather than relying on older pricing or feature comparisons, as Google’s subscriptions and model limits can change.
Also read: How to Use Gemini Floating Bubble on Android While Using Other Apps
Final Thoughts
Google’s upcoming Gemini changes put greater emphasis on the difference between free access and paid AI subscriptions. Free users are expected to lose Flash and Pro, while AI Plus subscribers will retain Flash but lose Pro access. AI Pro and AI Ultra will continue to include the full range of Flash-Lite, Flash, and Pro models.
The planned arrival of Deep Think on AI Pro could help balance these restrictions by making advanced reasoning available to more paying subscribers. New thinking-effort controls may also give users greater control over how Gemini handles different tasks.
With Gemini 4 Argon moving through its restricted rollout, Google’s AI plans are continuing to evolve. For users, the most important thing is to understand which features their subscription includes and choose a plan based on how they actually use Gemini.