Hello @chapperscobbler
You're correct - in Microsoft Fabric, the Compare runs view is currently limited to 100 runs at a time. This is a UI limitation, not a restriction on how many runs an experiment can hold, and it applies consistently across tenants and capacity SKUs.
A common workaround is to reduce the number of runs you bring into the comparison by filtering first. In the experiment UI, you can filter by metrics (for example, only runs above a certain score), parameters, tags, or time window, so that the “top” or most relevant runs fall within the 100‑run comparison limit. This is the intended usage pattern in Fabric and keeps the comparison grid focused and performant.
For larger-scale analysis, the recommended approach is to use MLflow programmatically in a Fabric notebook. By querying runs through the MLflow API, you can retrieve hundreds or thousands of runs, load them into pandas or Spark, and perform your own comparisons, rankings, or visualisations (for example, sorting by metric or plotting hyperparameter sweeps). This bypasses the UI limit entirely while remaining fully supported within the Fabric Data Science experience.
I trust this will be helpful. If you found this guidance useful, you are welcome to acknowledge with a Kudos or by marking it as a Solution.
