I am seeing a new error in an agent that was previously working as expected. In test mode I am seeing:
The request is resulting in too much data to handle, please evaluate the amount of data being returned by your actions.
Error Code: OpenAIModelTokenLimit
The error above started showing up in a very simple agent on Monday Apr 20, 2026. Prior to this date, we used the agent without issue and never ran into this error in our testing. I re-ran test scenarios that had previously passed and now they are throwing this new error. This is a simple agent with a single knowledge source. The source is a SharePoint document library. I added this knowledge source by choosing SharePoint and using the UI to navigate to the proper site and the proper library. I did not select individual files and the library only contains approx 26 files related to general IT questions. The files themselves are very small as well. Again, all of this was working without issue a few weeks ago. Thanks for any insight you might have!
It seems that the error is caused by the agent retrieving too much content from your SharePoint knowledge source, exceeding the model’s token limit.
This likely started due to backend changes in retrieval behavior.
The best fix is to reduce the scope of the knowledge source (select specific files instead of the entire library), optimize document size, and ensure queries return smaller, more focused content.
Was this reply helpful?YesNo
Under review
Thank you for your reply! To ensure a great experience for everyone, your content is awaiting approval by our Community Managers. Please check back later.