With only 5 to 10 users overall, you will focus on qualitative analysis. The sample will be too small to do any kind of meaningful quantitative analysis.
To gather qualitative feedback, it is often useful to show 2 different options, and ask the testers to compare. You will get more in-depth and valuable feedback. This approach is even more valid here, as it's not 2 alternative designs for the same device, but for 2 different devices. You will be able to assess the consistency of the UI across platforms, and the transfer of familiarity from one to another. I would suggest to vary the order, start with mobile then desktop for the 1st group, and vice versa for the 2nd group (assuming that your users can start their journey with either device).
If you wanted to do a quantitative analysis with a larger sample, it would be preferable to have 2 separate groups, one for each device / option (and assign testers randomly to either group). If users test both devices, you will need to differentiate in the results tests on mobile seen first versus tests on mobile second, to see if the order had any impact.