​
Create a comparison dashboard

1

Name your report

Name your comparison report something descriptive (e.g., “Co-pilot Nov Updates Comparison”)

2

Select runs to compare

Pick the runs you want to compare by clicking the add button next to each one

3

Set a base run

You can set any run as your base run to compare others against

4

Filter runs

Use the search bar and filters to find specific runs

5

Create dashboard

Click “Create dashboard” and you’re all set

​
Understand your comparison report

You’ll see several key metrics and visualizations:

  • Summary by Evaluator
  • Cost by Prompt
  • Token usage
  • Latency metrics

If you’ve set a base run, you’ll see how metrics change compared to that baseline.

​
Update your report

Hover over the report title and click “Edit” to add new runs, remove existing ones, or change your base run.

​
Share your report

Just click the “Share report” button at the top of the page to share with your team.