Profiler: Comparing Models Beyond Temp & Max Response Size

In the profiler, I want to compare outputs from the same model but with different parameters. For most models, I can tweak Temperature and Max Response Size, but is it possible to go deeper than that?

For example:

  1. Toggle Reasoning (if available) on/off, or set a max reasoning size.

  2. For GPT-5 specifically: adjust Reasoning Effort (low/medium/high).

Is there a way to do this in the profiler, or am I stuck with just temp and max response size for now?

Hi @bkermen,

Temperature and Max Response Size are the only parameters you can currently adjust in the Profiler.

Thank you for the suggestion! I’ve moved your post to the Feature Requests category.

1 Like