Discussions

Are rate limits per model or are cummulative?

In other words, if an user is iterating multiple models, would reaching rate limit on one model automatically block access to other models or rate limits are being monitored independently for each model separately, so if blocked for one model, you can switch to another one?