Hi everyone,
I’ve been trying to get Gemini 2.0 running but keep hitting walls. Here’s a summary of my attempts and frustrations:
- Google AI Studio → Get API Key Failure
- Whenever I try to get an API key, it says “cannot create a new project.”
- I had 7 projects, deleted 3, and now I’m at 4—but it still refuses to create a key.
- Generative Language API Only Shows 1.5
- Even though I want Gemini 2.0, it defaults to Gemini 1.5 Pro whenever I make calls.
- No matter how I specify the model (e.g.,
gemini-2.0-flash
,gemini-2.0-flash-001
, etc.), it doesn’t work.
- Vertex AI Setup & SDK Attempts
- I installed the Google Cloud SDK, set up
gcloud auth application-default login
, verified my region isus-central1
. - Despite all that, Vertex AI doesn’t recognize Gemini 2.0 at all—any call returns “Unknown model” or “404 Not Found.”
- I’ve tried using the Vertex AI Python SDK (with
ChatModel.from_pretrained
) and direct REST calls with the chat endpoint—no luck.
My end goal is pretty simple: I want to summarize some PDFs with Gemini 2.0 in batch (instead of manually uploading them to the AI Studio interface). But every route I’ve tried either fails to generate a new API key or only uses the older Gemini 1.5 model. I’m stuck.
Has anyone faced similar issues or found a workaround? Do I need special access or a different setup for Gemini 2.0? I’d appreciate any pointers—this has been a real headache!
Thanks in advance!