Input Slicer not functioning properly

Hi powerbi_user_01​,

The Desktop versus Service difference in your screenshots is interesting, but I would try one test before treating this as a general Input Slicer failure.

The current Input Slicer documentation explicitly supports operators such as Contains all, Contains any, Starts with any and Is any in both Desktop and the Power BI Service, but it does not currently document * wildcard syntax for text fields.

For your example, try setting the operator to Contains all and entering coupling and .75 as two separate values rather than using coupling*.75*.

If that works correctly in the Service, I would narrow this down to a difference in wildcard parsing rather than the Input Slicer itself. If Contains all still returns different rows between Desktop and Service using the same published model, that would be much stronger evidence of a Service-side regression.

The Input Slicer is now GA, and Microsoft has had some previous known Service-side issues with the visual, but I cannot currently find a published known issue that specifically matches this Desktop/Service filtering difference.

Since you are already on the August 2026 Desktop build, I would also create a very small repro page with just the Description column, Input Slicer and table visual, publish that, and compare the same Contains all test in both environments. That should tell us whether the issue is the report/model or the Service rendering/filtering path.