Vertex AI API quotas. Documentation discrepancy.

[yehorh]

Good afternoon,

According to the documentation at https://cloud.google.com/vertex-ai/generative-ai/docs/quotas#quotas_by_region_and_model, the base model gemini-1.0-pro in the Iowa region (us-central1) should have available quotas of 300 rpm. However, in reality, the limit is set to 5 rpm.

I thought it might be because it's a free account, so I upgraded to a paid account, but nothing changed. Following a recommendation, I made a manual payment hoping that this would make the quotas align with the documentation, but again, nothing changed.

What could be the issue and who should I contact? The sales team said they couldn't help and advised contacting support, but you can only subscribe to support if you have an organization.

[yehorh]

I verified my business email and received a $100 bonus.
However, I still can't increase my quotas.

I think if I buy payment support, nothing will change.
It's confusing.

[lsolatorio]

Hi @yehorh, 

Thank you for joining our community.

I understand it's frustrating when the documented quota for the Gemini 1.0 Pro doesn't match what you see in the UI. It's possible you're currently receiving the quota for a different model or version.

I've checked for known Vertex AI issues, but none seem directly related. Before considering paid support, you could submit this as a quota discrepancy issue to the Vertex AI team. Their engineers might be able to shed light on the cause and potential solutions.

I hope this helps.

[yehorh]

I tried contacting support, attended Google I/O, tried recreating projects, and reached out to the sales team, but it was all useless.

Done.