NotFound: 404 models/imagen-3.0-generate-001 is not found for API version v1beta, or is not supported for predict. Call ListModels to see the list of available models and their supported methods.
Even after completely following Documentation I am getting this error.
Note the warnings at the top of the page:
Preview: Imagen 3 in the Gemini API is available as an early access release in private preview. It’s not yet generally available.
If you’re in the early access program, you should have access with the account and project you’ve requested. If you have issues - you should message your contact at Google who granted you access.
If you’re not in the program - you won’t have access yet.
I posted about this in stack overflow. Here’s a copy of my sleuthing:
Got the same problem. If you run
!pip install -U git+https://github.com/google-gemini/generative-ai-python@imagen
import google.generativeai as genai
import pprint
for model in genai.list_models(page_size=100): # page_size can be any number
pprint.pprint(model.name)
Available Models
Instead of seeing the Imagen 3 model (“imagen-3.0-generate-001”) or any image generation models, here’s what it returns, or at least what I get:
Problem:
According to Google’s blog post, there is a waitlist for accessing Imagen 3. To join, you need to fill out the “Labs.google Trusted Tester Waitlist,” which requires providing credentials—a process that feels like getting military clearance.
Has anyone successfully accessed Imagen 3 through the API? Any tips or alternative steps to get this model up and running?
Welcome to the forums!
People who have access as part of a Trusted Tester program are, generally, under NDA and not allowed to discuss it.
There is no workaround. Access is linked to accounts and/or projects.
I have access to Imagen 3 through the Vertex AI API, and I’ve spoken about this, but it sounds like you’re asking more about how to get access than anything about the API itself?
Picking up the topic again.
I also have access to the Trusted Tester program, and despite having had a working implementation using Vertex AI it seems not to work anymore. The model names are showing up on Vertex AI and I can use Imagen 3 (Fast).
Edit: After using the correct API version I have it running in Vertex AI again.
Now, following the documentation as mentioned above for Google AI / API using the latest imagen branch of generative-ai on Python 3.12 I’m getting same HTTP 404 in response.
google.api_core.exceptions.NotFound: 404 models/imagen-3.0-generate-001 is not found for API version v1beta, or is not supported for predict. Call ListModels to see the list
of available models and their supported methods.
Can anyone confirm whether this is actually working?
Having access via Vertex doesn’t necessarily guarantee access through the AI Studio API. (The generative-ai library, probably.)
Again, people who have access through the AI Studio API are likely unable to tell you if it works or not, since they are under NDA.
Hi Allan,
yeah, I figured.
I had an exchange with Frank G. via email recently and yes, despite being the same model there are two independent EAP for Google AI and Vertex AI. Having access to the documentation with the sample code in Python gave me the idea to adopt it into my Gemini SDK for .NET as an “early adopter”. Boy, what did I think? 
Anyway, I have the implementation in place. Since weeks, perhaps even months, it works in Vertex AI and it won’t be an issue to test (and adjust, if necessary) as soon as Imagen on Google AI has been made available.
It would have been nice and practical to write another sample application in .NET that shows the capabilities of Google’s AI models. And of course, provide feedback to the team.
Cheers, JoKi