Is there a limit to users accessing assistants?

API
,
1

I am trying to build an application that multiple users (1000+) can access at the same time. The way I am approaching this is to create a single assistant that has its instructions. Then every time a user wants to use it and get info or chat with it, I create a thread for that particular session. This would mean, at any given time, there could be 1000 threads that a single assistant would be working with and more as users grow.

I have looked at the rate limit documentation, and while my organization will increase the tier as we grow, my main point of concern is would a single assistant be enough to handle this? Or should I create multiple assistant with some kind of load balancing logic of my own when calling the APIs?

2

It simply cannot be deployed to users in its current state, unless you are ready to bill them $1+ per input just like OpenAI will bill you when the assistant goes out of control or is made to do exactly what the user wants recursively, with a price you can’t programmatically retrieve. And that you can’t make enough requests to support that user base:

During this beta, there are several known limitations we are looking to address:

  • 60 req/min limit at the user account level.
3

Hey I’m looking at doint the exact same: same assistant , but different threads. Is there a workaround by now? Because I have been trying with only 5 users and it wasn’t working very well parallelly