Hi,
Google Gemini API for 1.5 Pro feels like a mess. We’re trying to use it in production, and we’re in the paid tier. We get about 50 requests per minute. I cannot figure out a way to increase the limits. The quota usage shows zero usage relating to Gemini.
How do we fix this? Why is there quota in the first place for paid users, shouldn’t we be able to scale automatically?
Welcome to the forums!
Certainly the AI Studio side of it is, I’ll agree with you there. I’ve had (and heard of) much better luck on the Vertex AI side of things.
There is a way to request a quota increase. See here for details.
Yup. Still no clue (and no response) about why this is happening. Pretty slipshod for a so-called production service.
There are good reasons for this for both Google and you.
Most Google cloud services have a quota. (For that matter, most API services from most companies have a quota.) This is to make sure you (or someone maliciously) doesn’t abuse the service.
Initially, as you’re starting out, that quota is set fairly low. This is to make sure you pay your use bill and you have reasonable controls over your API usage. And, since you’re just starting, a low usage is expected. But as you’re doing more, as you can demonstrate you’ve having that need, and as you can show you’re paying your bills, you can request a higher limit that should fairly routinely be approved.
The problem is - with the AI Studio Gemini API right now… that appears to be missing.