I’ve been looking high and low and cannot find RPM and TPM limits for 2.0 pro experimental.
We’ve been using 1.5 pro for the past 6 months and so tired of some of the responses that finally cut bait and started using 2.0 flash and it does some crazy things (for example, we create some glossaries out of long context documents and it produces the same term over and over and over again).
Wanted to look at 2.0-pro-exp but run into errors and lack of documentation on this model is frustrating. I get is an experimental model, but can’t even find pricing. Is it free?
It is currently free as it is experimental and not GA yet. If you put your cursor over the model where you select it in AI studio, you will see the information you are looking for. See screenshot for reference (including RPM and daily limits):
Thank you, and yes, I did see that, but we are testing via the API. If is free why the 1000 RPM conflicting with the “free” 2 RPM and 50 req/day below? No worries if don’t know.
I can’t nail this down so we will be reverting back to 1.5 pro unfortunately even though was hoping to leverage new models for some long context needs we have.
Not sure why you mentioned API as if your thoughts were these numbers do not apply to API use. These numbers are specially addressing the limits and use case for API use.
As for the top and bottom numbers, across all models, the bottom number is what you get in the free tier. The top number is what you will see in the paid tier. This is obviously confusing for models that are not GA yet but still experimental since there is no paid tier yet, but I believe they post it for planning purposes for those of us planning on selecting the right model in the future. For example, the paid tier for the thinking model shows as 10RPM. This number is obviously very low due to the resources it will consume iterating through its thoughts process. So knowing that very limiting number up front will help those currently writing applications testing with various experimental models so we can focus on the right ones.
Note that 2.0 flash is GA, you can use that and it is better and faster then 1.5 pro. You just can’t use 2.0 pro outside of the limits due to it not being GA yet.
I want to know if there is a limit like 2 RPM 50 req/day on the website https://aistudio.google.com/. If so, this is obviously not enough for our normal use. How to solve this problem?
Those are 1.5 pro and 2.0 pro Limits for free tier. Try 2.0 flash. 2.0 flash is better and faster then 1.5 pro and allows 15RPM and 1,500RPD. If you really need to test 2.0 pro, and those limits are not enough then sorry bud. I wish they were higher also although for most of my dev tests they are enough. But in all honesty, 2.0 flash is pretty amazing. I don’t see the real need for 2.0 pro yet for most tasks I have and I am doing some very complex multimodel reasoning.
