Hello NVIDIA Support Team,
I am writing to request a rate limit increase for my NVIDIA NIM API account.
I’m a single videogame developer that is experimenting with the platform in his own free time. My main interest at the moment is using agentic AI to create a narration toolbox for a person to play with.
I’m interested into test various models and undestand the differences. I mostly work on my local models, but something important I want to test is comparing how local models (like gemma 31b or qwen 27b) performs compared to larger, cloud base model like Kimi 2.6 or MiniMax2.7.
I tend to hit this limit from time to time, however it seems the limit then is enforced for much longer than a minute which is a problem.
I wouldn’t mind to have the 40 requests limit with a true 60 second cool down, my system can manage that. However with 200 requests per minute this will never happen as the system only spawn between 50 and 80 request per minutes at worst, and usually has various minutes in between these burst.
The way that this works is like this:
AI Model write various files(3/4) to create permant memory about the state of the “world” it generates. Then it narrates to the player and wait for an action.
Anywya, thanks and I hope you can accomodate my request :)