TL;DR
I wanted to rule out the possibility that the unusually low AI Studio limits were specific to my existing Pro account, its history, or my usual long-context workload.
So I tested two completely separate Google accounts:
Account A: no Google AI subscription
Account B: separately subscribed to Google AI Pro, with AI Studio visibly recognizing it as PRO
I used the same kind of short, simple prompts on both accounts and avoided my normal long-context RP workload.
Result: both accounts reached the same observed practical generation limit.
The non-subscribed account explicitly displayed:
“Upgrade to unlock more”
“Access higher limits, Pro models, and more.”
I then subscribed the other account to Pro.
The subscription succeeded.
AI Studio displayed PRO.
The observed generation allowance did not increase.
This does not prove that Google officially assigns identical Free and Pro quotas. There may be an entitlement delay, provisioning problem, quota-assignment issue, dynamic quota mechanism, or another backend explanation.
But it does establish one narrow observation:
> In my test, purchasing Google AI Pro produced no observable increase in the included AI Studio generation limit.
So I would like to know whether this is expected behavior or a malfunction.
-–
Why I ran this test
In my earlier report, I described a dramatic reduction in the practical number of AI Studio generations available on my existing Google AI Pro account.
There were several reasonable explanations.
Perhaps my existing account had an entitlement or provisioning problem.
Perhaps my very long RP contexts were consuming some hidden compute-based allowance faster than I realized.
Perhaps account history mattered.
So I removed those variables as much as I reasonably could.
I did not use my existing Pro account.
I did not use a long-running RP conversation.
I did not use hundreds of thousands of tokens of context.
Instead, I used two separate accounts and short, simple prompts.
The question was intentionally narrow:
Does subscribing a separate account to Google AI Pro produce an observable increase in the included AI Studio generation allowance compared with a non-subscribed account?
In this test, it did not.
-–
Account A: no Google AI subscription
The first account had no Google AI subscription.
AI Studio displayed:
“Upgrade to unlock more”
and:
“Access higher limits, Pro models, and more.”
I used short prompts until the included-generation limit was reached and recorded the result.
That account remained unsubscribed.
-–
Account B: newly subscribed to Google AI Pro
For the second test, I used a completely different Google account and subscribed it to Google AI Pro.
The subscription completed successfully.
AI Studio visibly recognized the account as:
PRO
I then repeated the same type of short-prompt testing.
No old RP history.
No massive context.
No comparison between a lightweight Free test and my normal heavy Pro workload.
Just a separate account, newly subscribed and recognized as Pro, under the same basic testing approach.
It reached the same observed generation limit as the non-subscribed account.
-–
What this does — and does not — establish
I want to keep the conclusion narrow.
This does not prove that Google’s internal Free and Pro quota values are officially identical.
I cannot see Google’s backend quota configuration, and the exact subscriber-allocation formula is not visible to users.
There are several possible explanations:
the Pro entitlement was not provisioned correctly;
there is a delay between the UI showing PRO and the higher backend allowance becoming active;
the account was assigned to the wrong quota bucket;
new accounts may have some additional restriction;
a dynamic or compute-based mechanism happened to produce the same result;
or this may actually be intended behavior.
I cannot determine which explanation is correct from the outside.
What I can say is simply:
> A non-subscribed account and a separately subscribed Pro account produced the same observed generation limit in my test.
That seems worth investigating.
-–
The sequence is the part I cannot reconcile
From the customer side, the sequence was very simple:
No subscription
→ AI Studio says “Upgrade to unlock more.”
→ AI Studio says “Access higher limits.”
→ I subscribe a separate account to Google AI Pro.
→ Payment succeeds.
→ AI Studio recognizes the account as PRO.
→ I repeat the test.
→ Observed limit: unchanged.
If the higher allowance takes time to propagate, that would explain it.
If provisioning failed, that would explain it.
If some other quota system is involved, that may explain it too.
But if this is working exactly as intended, then the obvious question is:
What did “Access higher limits” actually unlock in this test?
-–
This is not explained simply by my long-context workload
My normal AI Studio workload does use very long contexts.
I understand that hundreds of thousands of input tokens require far more compute than a trivial prompt.
That matters enormously for paid API usage.
But that is precisely why I did not use that workload for this comparison.
This test was not asking:
“Why does my long-form RP consume so much compute?”
It was asking:
“Does a Pro subscription produce a higher observable included AI Studio allowance under a simple short-prompt test?”
I could not observe one.
So long context may still explain part of my original account experience, but it does not explain this result.
-–
This is about the AI Studio subscriber allowance, not Gemini API Paid Tier
I also want to avoid mixing two separate systems.
I am talking specifically about the included AI Studio usage associated with Google AI Pro.
I am not talking about:
Gemini API Paid Tier throughput
metered API billing
paid API keys
API production capacity
I am not arguing that Google AI Pro should make paid Gemini API calls free.
That is a separate product and billing system.
This test concerns the benefit shown directly inside AI Studio.
The non-subscribed account was told that upgrading would provide higher limits.
The other account was upgraded and recognized as PRO.
The observed generation limit remained the same.
-–
Why I think this deserves attention
Other users on this forum are already reporting sharply reduced AI Studio availability.
Interestingly, the reported numbers are not even consistent.
Some users report reaching the limit after only a handful of generations.
Others report around ten.
My existing Pro account has often been closer to 15–20.
There have also been reports from higher subscription tiers of unexpectedly low limits.
That inconsistency is one reason I have wondered whether this could involve entitlement, provisioning, or quota-assignment problems rather than simply one new fixed limit.
This test does not prove that theory.
But it does make it harder for me to attribute the problem solely to my original account or my long-context workload.
I changed the account.
I changed the subscription state.
The result did not change.
-–
What I would like clarified
I think the questions are now fairly specific:
1. Should a newly subscribed Google AI Pro account receive a higher included AI Studio allowance than a non-subscribed account?
2. Should that allowance be active once AI Studio visibly recognizes the account as PRO?
3. Can the UI show PRO before the corresponding quota entitlement has been provisioned?
4. Are there currently known issues with AI Pro entitlement, quota assignment, or subscriber allowances in AI Studio?
5. Can Free and Pro accounts legitimately reach the same observable generation limit under otherwise similar short-prompt testing?
6. If this behavior is expected, what does “Access higher limits” mean in practical terms?
I am not asking Google to publish every internal quota formula.
I am not asking for a permanently fixed request count.
I am asking something much simpler:
Was the result I reproduced expected, or was something not applied correctly?
-–
Final result
Without interpretation, my test was:
Account A: no Google AI subscription
Account B: newly subscribed to Google AI Pro
AI Studio status for Account B: PRO
Workload: short, simple prompts
Observed practical generation limit: identical
The subscription status changed.
The observed allowance did not.
If this is a provisioning or entitlement problem, I hope it can be identified.
If another technical mechanism explains it, I would genuinely like to understand it.
If this is intentional, then “Access higher limits” needs a clearer practical definition.
-–
P.S.
One aspect I do find disappointing is the lack of visible staff acknowledgement so far.
There are now multiple threads and many replies from users independently reporting the same broad pattern: a sudden and substantial reduction in usable AI Studio generations, even though the exact numbers vary considerably between accounts.
I understand that forum moderators, support staff, and the teams responsible for AI Studio quotas are probably separate groups, and I do not expect an immediate technical diagnosis.
But with this many paying users independently reporting a sudden change, even a brief acknowledgement that the issue is being examined would be useful.
At this point, the silence is becoming almost as noticeable as the quota issue itself.
[Free Quota, each model around 10]
[After Subscribe, Other One, Same Quota]