Human 'peak' is a range, rather than a single rank. It is a different range for each ability. The human peak in Intellect is not the same as the human peak for Stamina or Strength. At least that is the thinking that informs the official GR npc's, yet when the ability benchmark chart was created in 2e, and transferred to 3e with a minor tweaking the idea inherent in that chart is to use a single benchmark for all abilities. The actual characters (At least those in print form by Green Ronin) suggest that chart is largely ignored and that instead a set of charts was mentally used in its place. A better replacement would be a chart that extends from -5 to 20 (or up to 30 potentially) and then columns for each ability or abilities that share that same benchmark set. For example, Intellect and Awareness might share the same though it need not too.
Next, the demonstrated result is that humans can and do function at meta or superhuman levels. Even real life humans demonstrate actual 'superhuman' capacities and depend on how one defines the term 'superhuman'. If we mean a tiger or ape is superhumanly strong as no human known can overpower them, yet often savants and other oddities of humanity can achieve truly 'superhuman' abilities, and IQ's that stretch in the 200's (perhaps up to 250 or greater).
Then there is the vaguery of the abilities themselves and exactly what they seem to really represent. One take is to see them as a collection of game functions like Agility, among other things as the ability to avoid ranged attacks, and Dexterity the prowess with ranged combat.
This is the one weak area in Mutants and Masterminds 3e in that it could have been clearly defined and a full set of ability benchmarks created instead of a single chart to be largely ignored as an afterthought. Unless, when you see a character who has a 15 it means that characters ability in that area is considered 'very high superhuman' in which case Slade the Terminator, human though modified with a super-soldier serum functions as a very high superhuman. One supposes than a non-human like Thor (or a human 'like' alien race) might function at a similar level as he is often referenced. Yet, in MnM the king is really the total modifier. If you had a 15 Fighting your actual final modifier is your true fighting ability and Fighting is more of a default level.
This is one area, abilities and benchmarks, that if there were ever a mutants and masterminds 4e I would like whomever designs it to give this real thought, rather than the confused effort presently at work.
Since skills are tied to abilities and total modifier is the true measure of your 'skill' having a high ability and less ranks but a high final modifier is really what should be measured. What limits your final bonus is your Power Level.
Brainaic for example is PL 13 and has a Technology +23, a 15 Intellect and is known to be of greater intelligence than Lex, and yet Lex has an 11, but a +24 Technology and simply makes up with more ranks. But does that make him 'better', yes but only by +1 and 'not as smart as brainaic'. So do you have skill benchmarks based on ranks in a skill or total modifier. I say total modifier since that is what really determines your overall ability. It means having a higher intellect just means your default skill modifier in all Knowledge or Technology related fields is higher on average. Brainaic is more capable on average but in one specific area, Technology, Lex has a slight advantage. I don't see Brainaic with any Science ranks but instead has Knowledge: Known Space at a rank Lex would have for Science. Then I see Brainaic 5 has having a 20 Science and Technology and only PL 10. He thus is also at PL max and placed at a few PL's below those other too but answers the books' designers idea that Brainaic 5 functions at a level slightly less than then.
All this leaves me with the impression that a single chart with multiple columns might have been best and then one column for Skill modifier (that includes any ability modifier) might be the 'best' or most solid approach. We shouldn't look at individual character right ups to see what benchmarks are but rather look at the benchmarks to know why a given character is that ability bonus or total skill bonus.
The 3e game books clearly has an approach but its not clearly spelled out. Instead you have to reverse engineer it and when doing so you'll see inconsistencies that could easily have been avoided if real thought had been taken at the start.