26.29 rounds to 26, not 27.
Nobody mentioned 27 though. 2.7 isn't 27.
And, although the wording clearly implies an absolute relationship, the correct relative formula would be 26.3/25.6=1.03 when significant digits are accommodated (which would be a 3% relative increase).
You're making the assumption that 26.3 and 25.6 are given with the full number of significant digits (which may not be the case), or that significant digits actually matter in a percentage figure (not an actual measurement) in popular scientific journalism. Get over yourself.
Well, at least you're in good company.
Someone's really salty to be shown wrong, eh?