[Bug] Paid Tier 1 account getting free_tier_requests limit:0 on image generation models (gemini-2.5-flash-image, gemini-3-pro-image-preview)

Gemini API
, , ,
1

  • AI Studio dashboard)

  • Project: Active billing account with valid payment method

  • Affected since: ~1 week ago (started around Feb 10, 2026)

Affected Models

  • gemini-2.5-flash-image → 429 error

  • gemini-3-pro-image-preview → 429 error

Working Models (same API key, same project)

  • gemini-2.5-flash (text/vision) → works perfectly

  • gemini-2.5-flash-lite → works perfectly

Exact Error Response

{
  "error": {
    "code": 429,
    "message": "You exceeded your current quota, please check your plan and billing details. For more information on this error, head to https://ai.google.dev/gemini-api/docs/rate-limits",
    "status": "RESOURCE_EXHAUSTED",
    "details": [
      {
        "@type": "type.googleapis.com/google.rpc.QuotaFailure",
        "violations": [
          {
            "quotaMetric": "generativelanguage.googleapis.com/generate_content_free_tier_requests",
            "quotaId": "GenerateContentRequestsPerModelPerMinutePerProjectPerRegion",
            "quotaDimensions": {
              "model": "gemini-2.5-flash-image",
              "location": "global"
            },
            "quotaValue": "0",
            "futureQuotaValue": "0"
          }
        ]
      }
    ]
  }
}

Key Observation

The quota metric says generate_content_free_tier_requests with limit: 0. My account is on Paid Tier 1, so it should be using generate_content_paid_tier_requests instead. Text/vision models correctly use paid tier quotas — only image generation models are affected.

What I’ve Tried

  1. Verified billing status in Google AI Studio → Shows “Paid tier 1” with available quota

  2. Created a brand new API key → Same 429 error

  3. Tested text models with same key → Work fine (confirms key and billing are valid)

  4. Waited several days → Issue persists

  5. Checked rate limit dashboard → Shows available RPM/TPM for these models

Impact

This is blocking our production application (e-commerce virtual try-on service) from generating images. The categorization and prompt optimization pipelines work fine with text models, but the actual image generation step fails 100% of the time.

1 Like