I have this thread and am quite impressed with some of the responses and would like to know which model choice it is using under the hood. I assume this is known but currently not exposed? Or is it exposed somewhere and I haven't found it.
#Is there a way to see what agent/model was used in an "auto" thread?
blazing sierra
maiden raftBOT
Hi @blazing sierra. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:
You’re right that Warp knows which concrete model was used behind an Auto selection, but the knowledge sources don’t clearly document a user-facing way to see that for an existing thread.
What is documented:
- Warp has multiple Auto modes that dynamically pick a model per request (Cost-efficient, Responsive, Genius). [Model choice]
- Warp uses a fallback chain when a selected model is unavailable, automatically switching to another model of similar tier. [Model choice; SWE-bench blog]
- A Warp team member mentions that the model you select is the one used, with fallbacks only on errors/outages, and that they’re “tracking making this more transparent,” but they don’t describe a specific UI element where you can inspect the final model per request. [Model confusion issue]
There are also GitHub comments from users referring to a “Credit details” window that shows which models were actually used on a request, but this is mentioned only in passing in an issue comment and not documented in the official docs. [Model confusion issue] Because it’s not in the docs, I can’t confidently describe how to open or use that view.
So, based on the available sources:
- Warp internally tracks which concrete model an Auto thread used.
- The official docs do not clearly document a reliable, supported way for you to inspect that per-thread/per-message yourself.
- Some UI for seeing model usage may exist (via credit details), but it’s not formally documented, so I can’t give exact steps.