Paid Tier 1 project: all models return "permission denied" / "internal error" — image and text generation both affected

Gemini API
1

Hi all,

Our Gemini API project has been failing on every request type for several days now, and I’d appreciate help narrowing down the cause since this doesn’t look like a normal quota or billing issue.

Project ID: gen-lang-client-09451360..

Billing: Tier 1, paid, CA$352.06 cap on the linked billing account. Confirmed active and correctly linked.

Actual spend: only $1.38 over the last 28 days, so this is not a spend-cap or overage issue.

API keys: 3 keys, all generated directly in AI Studio (Aug 8, 2026), all Tier 1 / Prepay.

What’s failing:

  • Nano Banana Pro (gemini-3-pro-image): fails with “permission denied” and “An internal error has occurred”
    • Nano Banana 2 (gemini-3.1-flash-image): same two errors, identical behavior
      • Gemini 2.5 Flash, plain text only (“Say the word test.”): same two errors
    • So this reproduces across three different models and across both image and plain text generation, which rules out a model-specific capacity ceiling in my mind. It looks account- or project-level.
  • I noticed the banner here about the June 19, 2026 change requiring restricted API keys. All three of our keys were generated in AI Studio, which the banner itself says are restricted to the Gemini API by default, so I’m not sure that’s the cause, but flagging it in case there’s a subtlety I’m missing.

Since there’s no ticket-based support for this tier, posting here per the forum guidance. Happy to provide request IDs or any other diagnostic info. Thanks for any help.

2

Hello @Survival_Tactix ,

We have recently rolled out a mandatory age verification initiative across signed-in customers globally, Accounts that have not completed explicit date-of-birth/identity verification should do so by navigating to Google Account Age Verification or verifying date of birth in Google Account Personal Info."

Also, disconnect any VPN, proxy, or iCloud Private Relay and access AI Studio directly. After completing these steps, please try again in a few hours, If you still face issues, please let us know.