Hi, I’m trying to generate images using Vertex AI Imagen (Model: imagen-3.0-generate-002),
but I keep getting a NOT_FOUND error, even though I am using the correct project, region, and publisher model path.
Error message: Publisher Model projects/(PII Removed by Staff)/locations/us-central1/publishers/google/models/imagen-3.0-generate-002not found
My Node.js code is calling the model via PredictionServiceClient:
const endpoint = `projects/${projectId}/locations/us-central1/publishers/google/models/(PII Removed by Staff);
const [response] = await client.predict({
endpoint,
instances: [
{
prompt: { text: prompt }
}
],
parameters: {
sampleCount: 1,
aspectRatio: "ASPECT_RATIO_1_1"
}
});
I also checked my gcloud settings:
gcloud config list
Core project: (PII Removed by Staff)
Region: us-central1
And Vertex AI model list:
gcloud ai models list --region=us-central1
Listed 0 items.
My questions:
-
Is the Imagen model imagen-3.0-generate-002 publicly available in all projects?
-
Do I need to enable additional APIs or access?
-
Do I need to apply for allowlist access for this Imagen model?
-
What is the correct model ID for the current public Imagen API?
Additional context
-
Service account authentication is working.
-
Permission errors (PERMISSION_DENIED) were resolved.
-
Now only the NOT_FOUND error occurs.
-
My goal is to generate images for a temporary prototype (1–2 weeks demo).
Any help would be greatly appreciated!