If you want the y-axis to look more uniform, consider changing the scale to a semi-logarithmic plot, or transforming your y data through a log transformation:

Option #1 - Change y-axis to log scale:

Simply obtain the current axes and set the YScale to log:

x = [0.1 1 10 100 1000 10000];
y = [2.66    24.58    81.72    87.90    87.99    88.00];
xplot = 1:numel(x); 
figure;
bar(xplot,y);
set(gca,'XTick', xplot); 
set(gca,'XTickLabel', x);
set(gca,'YScale','log'); %// NEW
%ylim([0 100]); %// Remove as we don't need it

This is what I get for the graph:

Be mindful that the y-axis doesn't change its data - only the way you present it. The ticks are placed in a semi-logarithmic basis where from the bottom, we go from 1, 2, 3, up to 10 (10^1), next we go 20, 30, 40, up to 100 (10^2), then we go 200, 300, 400, up to 1000 (10^3) and so forth. You can certainly see a pattern in the ticks, as we are increasing on a semi-logarithmic basis.

Option #2 - Transform your y-data to log:

Just take your y data and run it through log:

x = [0.1 1 10 100 1000 10000];
y = [2.66    24.58    81.72    87.90    87.99    88.00];
xplot = 1:numel(x); 
figure;
bar(xplot,log(y)); %// NEW
set(gca,'XTick', xplot); 
set(gca,'XTickLabel', x);
%ylim([0 100]); %// Remove as we don't need it

This is what we get:

Be mindful that the y-axis has changed its units. You can leave a disclaimer saying that your data was log transformed.