In the context of web display, '100% crop' often means to display a section of the photo which easily fits within anyone's monitor (no matter how low in total resolution they have...like 1300 x 768!) where there is NO SOFTWARE decimation being done in their PC/phone to 'fit' your pixels within the resolution limitations of their monitor...that is, 100% of your image's pixels are displayed 1:1 on the viewer's monitor pixels.
To explain further, an R6 sensor has 3684 pixels vertically, but a 4k monitor only has 2160 pixels, so it is impossible to display the R6 image 'at 100%', at best (and ignoring any pixels being used by a photo application for user control display) the R6 image is on the monitor at 58% of actual size and 42% of them are 'decimated' by the display software. If you wanted to show every pixel in the photo (and not have software decimation of pixels), you would need to take a section of the image which occupies (for example) 1500H x 1000V pixels and no one's PC or tablet or phone would use software decimation in order to display the entire image with all its pixels on the device's monitor.
When taking a 100% crop from two cameras and comparing photo quality, one has to be very careful to understand that cropping is NOT simply removal of a section of pixels, one is also removing some of the resolution of the original lens as well!
Let us assume two cameras with same overall size sensor, but very different (mythical cameras) resolution... 2000 pixels vertical in Camera 1, and 20000 pixels in Camera 2. Let us begin by assuming the lens itself delivers 2000 line-pairs of detail resolution vertically.
By taking a 2000 pixel section from Camera 1, and comparing a 2000 pixel section from Camera 2, it seems to be comparing apples to apples, but it is NOT.
I am taking a 2000 line-pair image from Camera 1, but comparing only 1/10 of the 2000 line-pairs from Camera 2...and Camera 2 looks far worse than Camera 1 even though its pixels better resolve what the lens delivers.
Ergo, comparing 2000 pxiels vertically from both cameras is an unfair comparison that makes Camera 1 seem better.