This limitation is because you are in a free trial. The model that powers the assistant has a low rate limit of requests per minute.
Unfortunately, this will prevent you from evaluating an assistant that makes multiple iterative calls. These iterations are often out of your control, made by internal functions such as for file retrieval.
The solution is to add a payment method and purchase a prepay credit for usage.
You can also code agent-like behavior yourself using the chat completions endpoint, and limit the rate at which API calls are made. This also can be more budget-friendly.