Google Vertex AI (generative AI) error
Vertex AI 403 PERMISSION_DENIED: what it means and how to fix it
403 PERMISSION_DENIEDRetry? NoLast verified 2 documented fixes2 sources
What it means
The client does not have sufficient permission to call the API.
Causes
- A required API is not enabled
- The service account lacks permission, for example to the Cloud Storage bucket hosting image or video inputs
How to fix it
- Fix 1.
Verify that all necessary APIs are enabled and that the service account has permission to access the selected service (see Access control).
- Fix 2.
Grant the per-product, per-project service account (P4SA) the permissions needed to access resources referenced in the input.
More Vertex AI errors
All 11 Vertex AI errors400 INVALID_ARGUMENT / FAILED_PRECONDITIONRetry? No401 UNAUTHENTICATEDRetry? No404 NOT_FOUNDRetry? No429 RESOURCE_EXHAUSTEDRetry? Yes429 Resource exhausted, please try again laterRetry? Yes429 Too many requests. Exceeded the Provisioned ThroughputRetry? Not stated499 CANCELLEDRetry? Not stated500 UNKNOWN / INTERNALRetry? Yes
HTTP 403 at other providers
Compare 403 across providers403 AccessDeniedExceptionRetry? NoAmazon Bedrock403 InvalidClientTokenIdRetry? Not statedAmazon Bedrock403 MPAgreementBeingCreatedRetry? YesAmazon Bedrock403 AWS Marketplace Agreement Failed within 15 minutesRetry? Not statedAmazon Bedrock403 AWS Marketplace Agreement Pending after 15 minutesRetry? YesAmazon Bedrock403 permission_errorRetry? Not statedAnthropic API403 Forbidden (network access denied)Retry? Not statedAzure OpenAI403 Forbidden (resource has been temporarily blocked)Retry? Not statedAzure OpenAI403 permission_deniedRetry? NoGemini API403 PERMISSION_DENIEDRetry? NoGemini API403 ForbiddenRetry? Not statedGroq API403 ForbiddenRetry? Not statedMistral API
FAQ
What does the Vertex AI error "403 PERMISSION_DENIED" mean?
The client does not have sufficient permission to call the API.
How do I fix Vertex AI "403 PERMISSION_DENIED"?
There are 2 documented fixes. Start with the first: Verify that all necessary APIs are enabled and that the service account has permission to access the selected service (see Access control).
Should I retry Vertex AI "403 PERMISSION_DENIED"?
No. The official docs treat this error as not retryable: fix the cause before sending the request again.
Which other AI APIs return HTTP 403?
Per their official error docs: Amazon Bedrock (403 AccessDeniedException; 403 InvalidClientTokenId; 403 MPAgreementBeingCreated; 403 AWS Marketplace Agreement Failed within 15 minutes; 403 AWS Marketplace Agreement Pending after 15 minutes), Anthropic API (403 permission_error), Azure OpenAI (403 Forbidden (network access denied); 403 Forbidden (resource has been temporarily blocked)), Gemini API (403 permission_denied; 403 PERMISSION_DENIED), Groq API (403 Forbidden), Mistral API (403 Forbidden), OpenAI API (403 Country, region, or territory not supported; 403 PermissionDeniedError), OpenRouter (403 permission_denied; 403 Budget limit exceeded; 403 content_policy_violation; 403 refusal), xAI API (403 Forbidden).
Sources
Last verified against 2 sources: Vertex AI's official docs and official SDK repositories.
Newsletter
Know when the APIs change
One email per week. New model releases, API changes, what broke, and how developers are working around it.