Just 5 requests per minute is very little, will this increase in the near future?
1 Like
Hi Ana, thanks for your post!
Which model are you referring to? Gemini 1.5 Pro has 2RPM at the moment and Gemini 1.0 Pro has 15.
We have a pay-as-you-go option launching very soon that provides higher RPMs 
5 Likes
I am a bit confused about the rate limits too, as far as I can see the limits for the latest Gemini models is:
Free tier: 2 RPM/50 RPD
Pay as you go: 10 RPM/ 2000 RPD
The same site also states that you can request higher rate limits, but by how much?
It’s also mentions that I can migrate to the Vertex AI platform on Google Cloud which may offer higher limits. But according to the following table, the rate limits are the same:
| Features | Google AI Gemini API | Google Cloud Vertex AI Gemini API |
|---|---|---|
| Latest Gemini models | Gemini Pro and Gemini Ultra | Gemini Pro and Gemini Ultra |
| Sign up | Google account | Google Cloud account (with terms agreement and billing) |
| Authentication | API key | Google Cloud service account |
| User interface playground | Google AI Studio | Vertex AI Studio |
| API & SDK | Python, Node.js, Android (Kotlin/Java), Swift, Go | SDK supports Python, Node.js, Java, Go |
| Free tier | Yes | $300 Google Cloud credit for new users |
| Quota (Request per minute) | 60 (can request increase) | Increase upon request (default: 60) |
| Enterprise support | No | Customer encryption key, Virtual private cloud, Data residency, Access transparency, Scalable infrastructure for application hosting, Databases and data storage |
| MLOps | No | Full MLOps on Vertex AI (Examples: model evaluation, Model Monitoring, Model Registry) |
(Ref: link)
2 Likes
If I’m not mistaken, until yesterday it was even less, it was 5 RPM on pay-as-you-go.
Another thing, the deadline for releasing pay-as-you-go access would be May 2nd, now it is May 14th.
In a production application this limit would certainly be a problem.
1 Like
I’m planning to use Pay-as-You-Go but 10 RPM would not be enough for my application, because for each user request I need to make 2 requests to Gemini to refine the response. This would support a maximum of 5 simultaneous users.
1 Like