Can't increase quotas for Vertex content generation?

[andreas_281921]

I currently have a limit of 5 on aiplatform.googleapis.com/generate_content_requests_per_minute_per_project_per_base_model.

I'd like to try to increase the limit but I don't seem to be able to find out how. The edit Quota screen tells me to go talk with sales and when I do so, sales tells me to reach out to technical support - which requires a support package. Is this really the only way to unlock the API to actually start building with it?

Can anyone point me in the right direction here?

Best
Andreas

[Poala_Tenorio]

Navigating through service limits and quota increases can indeed be a bit convoluted, but there are steps you can take to try to increase your quota without necessarily needing to purchase a support package immediately. Here's a general approach you can take:

Follow the documented process for requesting a quota increase. This often involves filling out a form or contacting support. Be sure to provide all necessary information, including your project details, current usage, and the requested increase.

Remember that the process and requirements for quota increases can vary between different platforms and services. It's essential to follow the specific guidelines provided by the AI Platform API documentation and reach out to their support channels for assistance if needed.

[andreas_281921]

Hi Poala_Tenorio,

I've tried this, multiple times. That's why I'm here.

The quota editing view does not allow to contact anyone. It doesn't show the form you're referring to. It solely says, Contact sales. Which I tried! I really wanted to use Gemini. The Sales team though quickly realises that you're actually asking for Quota and then they say: "I'll give you the email from the technical support team. I'll email it to you." And then they close the chat. And the email you get refers to the quota documentation you've just linked to as well. I've been through this twice.

It is, as if Google doesn't want people to use Gemini and the approach to customer support is just mind-blowingly unhelpful. Sorry, this was just such a frustrating experience.

I've by now decided to go back to OpenAI. I seriously tried using Vertex but I want to build and not spend my time trying to hack the process-machinery.