Hello,
I’m hoping to get some help with a persistent issue trying to access Gemini models through the Vertex AI API. No matter which region or model I try, I consistently get a 404 Not Found or project does not have access error.
Test Goal: I am trying to call the Gemini model from a backend application hosted on Databutton to perform text extraction.
The Error: 404 Publisher Model projects/projectapp-49e4b/locations/us-central1/publishers/google/models/gemini-pro was not found or your project does not have access to it.
Troubleshooting Steps Already Taken:
- Vertex AI API is Enabled: I have confirmed in the Google Cloud Console that the “Vertex AI API” is enabled for this project.
- Billing is Active: I have confirmed there is an active and linked billing account for this project.
- Correct IAM Permissions: I created a dedicated service account with the “Vertex AI User” role. The application is using a valid JSON key from this service account to authenticate. The authentication itself seems to be working, as the error is coming from the API after authentication.
- Regions Tried: us-central1 and asia-east2 (both fail with the same error. My preferred region is asia-east2 for latency, but for troubleshooting, I also tested in us-central1 and received the exact same error.)
- Models Tried: gemini-1.0-pro and gemini-pro (both fail)
I am not sure what to check next. It seems like a project-level permission or configuration issue that I’m not seeing. Has anyone encountered this before or have any ideas on what else I could check or may have missed? Any help would be appreciated.
Thank you!