Examples of usage `generateImages` `editImage`?

[dantaeusb]

Since Imagen endpoints are gone from "Gemini Enterprise Agent Platform", there's literally not a single model that can work with these methods.

Documentation still says:

const response = await client.models.generateImages({
 model: 'imagen-4.0-generate-001',
 prompt: 'Robot holding a red skateboard',
 config: {
   numberOfImages: 1,
   includeRaiReason: true,
 },
});

but Vertex AI (now Gemini Enterprise Agent Platform) has no such models, they're gone: https://docs.cloud.google.com/gemini-enterprise-agent-platform/resources/locations.
The only model that have image generation capabilities are vllms that are available through generate/streamContent, not generateImages / editImage. Is there any use for those methods? Why are they not deprecated?

I am genuinely puzzled because some partner providers give access to Gemini family models (Nano Banana), which do seem to work with mask inpaining over OpenAI compatible APIs, but I could not find a way to make it work directly. Is there something I am missing?

[Varun-S10]

Hi @dantaeusb, Thank you for bringing this to our attention. We truly appreciate you flagging this issue. We will file a bug internally.

In the meantime:

• For Imagen generation on Vertex AI, please use imagen-3.0-generate-002 (in supported regions like us-central1).

[dantaeusb]

Yes @Varun-S10 I tried several options by recommendation from Gemini, none work, I have the following errors:

Publisher model `projects/.../locations/us-central1/publishers/google/models/imagen-3.0-generate-002` was not found or your project does not have access to it.

I have deduced that Imagen is simply no longer available (as it's no longer listed in models list in the Agent Platform docs), hence the question.
Agent Platform API, Gemini for Google Cloud, Gemini API, all enabled in the project.

P.S. Also it is likely that partner providers do not use image edit APIs or Imagen either, it was just coincidence that Gemini models were returning images modified by given mask, I believe mask is simply discarded. When trying to prompt Gemini directly to edit by mask, it often just adds mask as white overlay on the image. Seems like mask guided inpainting is simply no longer available at Agent Platform API.