404 Publisher Model `projects/project/locations/us-central1/publishers/google/models/text-bison@00

[anekanttru]

Hi,

I am creating a model to generate insights from BigQuery Ga4 tables using models/text-bison@001`, I have cross-checked the access and everything, but I am still getting the error.

404 Publisher Model `projects/ga-4-test/locations/us-central1/publishers/google/models/text-bison@001` was not found, or your project does not have access to it. Please ensure you are using a valid model version.

I have also tried multiple version but so far no luck, all the API are enabled and vertex is running in the project itself.

[dawnberdan]

Hi @anekanttru,

Welcome to Google Cloud Community!

The sample you used here uses the model “text-bison@001”. This model has been discontinued so that may be the reason you’re getting a 404 error.

You need to migrate to a newer, supported model. As of now, “gemini-2.0-flash” or other “gemini” models are the recommended replacements for text generation tasks. You can find the latest supported models and their lifecycle in this documentation.

When you switch, remember that the API call format and potentially the input/output schema might differ slightly between models. You can also double check the exact model ID you are using in your code. It should align with the current, supported model IDs from the Vertex AI Model Garden.

If the migration doesn’t solve it, re-confirm permissions. You can find information on managing project permissions for Google Vertex AI here and details on access control with IAM. If you're looking for specific IAM permissions required for different Vertex AI, check this document.

Was this helpful? If so, please accept this answer as “Solution”. If you need additional assistance, reply here within 2 business days and I’ll be happy to help.

[anekanttru]

Hey @dawnberdan,

Thank you for your response, I tried to use the updated model, but I am still getting the same error.

[dawnberdan]

Hi @anekanttru,

Since you've already switched to a supported model (such as gemini-1.5 Pro) and the issue persists, here are a few things to double-check:

1. Model ID: Ensure the model ID you're using is exactly correct. For example, it should be something like:

projects/PROJECT_ID/locations/LOCATION/publishers/google/models/gemini-1.5-pro

Replace PROJECT_ID and LOCATION (e.g., us-central1) with your actual project info.

2. API Endpoint: Confirm that the endpoint you're calling matches the correct region for the model. Some models are only available in certain regions.

3. Permissions: Double-check that your service account or user has the necessary IAM roles to access the Vertex AI model. At a minimum, you'll need:

• Vertex AI User (roles/aiplatform.user)
• Possibly Viewer (roles/viewer) depending on what resources you're accessing

4. Model Availability: If you’re using a custom model or a version that may not be deployed yet, ensure that the model is in the “Ready” state in the Vertex AI console

Hope it helps!