Fix Vertex AI 404 Model Not Found Error

Symptoms

1. Calling the Vertex AI or Gemini API results in a 404 Not Found error.
2. The error message looks like this:

GoogleApiError: Publisher Model 
`projects/[GCP_PROJECT]/locations/[GCP_LOCATION]/publishers/google/models/gemini-1.0-pro`
was not found or your project does not have access to it.

Note: [GCP_PROJECT] and [GCP_LOCATION] vary depending on your configuration.

In this example, I’m using gemini-1.0-pro, but the same applies to any relevant model. I’ll explain why in a moment.

Why It Happens and How to Fix It

A 404 error usually means that the model you’re trying to use either doesn’t exist in that region, was deprecated, or your project doesn’t have access.

1. Check the Official Models List

Visit the Google Models List:

• Confirm the model ID you’re using (ex: gemini-2.5-pro).
• Verify the supported regions for that model.
• If you’re using an unavailable model or region, Vertex AI will return a 404.

2. Verify That the Model Isn’t Retired

See the Retired Models List:

• If the model is listed here, it’s no longer accessible.
• Update your code to use the latest supported version.

3. Check Permissions

Make sure your account / service account has the proper IAM role:

• Required role: Vertex AI User (roles/aiplatform.user)
• This is needed to run prompts in Vertex AI Studio.

Without this permission, your project won’t be able to access Vertex AI models even if they exist.

Inspecting a Model Page

Let’s look at a real example: Gemini 2.5 Pro (as of November 2, 2025).

• Gemini 2.5 Pro Model Page

Here you can see that the Model ID is gemini-2.5-pro. That’s exactly what you must use in your API calls / code.

These are the supported regions for gemini-2.5-pro. If you try to call the model from an unsupported region, you’ll get a 404.

TL;DR

Last but not least, you should take a look at the page Which Gemini model should I migrate to to help you choose the upgraded model that best fits your needs.

By then, you should be good. 🫡