403 PERMISSION_DENIED on generateContent while models.list and free-tier quota work

Hello Google Gemini API team,

I am experiencing a persistent 403 PERMISSION_DENIED error with the Gemini Developer API.

Project:
Default Gemini Project

Project ID:
gen-lang-client-05339590..

Model:
gemini-2.5-flash-lite

What I have verified:

  • Gemini API (generativelanguage.googleapis.com) is enabled.
  • The project is active.
  • There are no warning banners or action-required notices in Google Cloud Console.
  • There are no warning notices in Google AI Studio.
  • A newly created API key was tested.
  • GET /v1beta/models succeeds with HTTP 200.
  • The models endpoint returns 56 models.
  • gemini-2.5-flash-lite is visible in the model list.
  • Free-tier quota is assigned and is not exhausted:
    • Requests per minute: 10, usage 1
    • Requests per day: 20, usage 1
    • Input tokens per minute: 250,000, usage approximately 14
  • Billing is not enabled because I am only testing the documented free tier.
  • POST generateContent still returns HTTP 403 PERMISSION_DENIED.
  • The same result occurred with a newly created API key.
  • I have not changed IAM roles, quota limits, Firebase settings, or billing configuration.

The API key is sent using the documented x-goog-api-key header.

Could you please check whether this project/account has an internal access restriction or automated flag, or whether this is the known Gemini API 403 issue?

There are no visible appeal or restriction banners in Cloud Console or AI Studio.

I can provide the project number privately if needed.

Thank you.