The method for determining the scores was used incorrectly
There are multiple methods in the official rules which you should not deviate from unless it is a conscious choice and was discussed with the group beforehand. Your question implies that neither is the case for you so sticking with the official rules is probably best.
Rolling for the scores is one method described and it uses 4d6 but discards the lowest result. The maximum achievable this way is 18, the same as with 3d6, the method used just makes the scores somewhat higher on average than with 3d6. The scores you describe indicate that 4d6 was used without dropping the lowest die. In which case base scores as high as 24 can result - plus ability score increases (ASIs) which explains the 25.
Another important point to notice is that player character ability scores are generally assumed to not surpass 20. That cap is not absolute but you can assume that it applies unless otherwise noted. See this question for some context.
The game balance between monsters and player characters in D&D 5E is rather loose and does favor the PCs anyway. All characters being strong is often not a problem and if it is, it is generally easy to fix by occasionally making encounters harder (adding more or stronger monsters).
The problem in your case is an unbalance between the different characters which does a) create an unfun play experience in most cases because the less powerful characters may feel “useless” for example and because it makes it harder for you to supply appropriate challenges, i.e., if all characters are very strong you can use harder encounters, but if one characters is much stronger, making appropriate encounters for them can easily be deadly for others.
There are not really any good options to fix this other than bringing all characters on a similar power level. The easiest option is to have the player in question make new stats. Tell them, assuming it was an honest mistake rather than any kind of ill intent, why it is important that characters have a similar power level. Also explain how to correctly determine stats and ask them to do that.
I have some more detailed explanations on balance below.
Discussing concerns of balance
The balance of monster stats in D&D 5E is pretty loose anyway and one or a two points of attack modifier won't change this too much. After all, the system assumes that no magic items are used because they're strictly optional but giving players weapons with a +1 or +2 bonus doesn't ruin the game. When I run games I also tend to be rather loose with the rules, e.g. I allow players a +3 ASI instead of +2/+1 to two different scores. I have found that neither has negatively influenced my game. That said, the difference from about 18 to 20 which is the typical maximum to 25 is equal to a +2 or +3 additional attack bonus which can have a significant impact. It is also better to stick closer to the rules the less experienced you are. The longer you run games the better your ability to foresee the impact of allowing or modifying certain things. You have to decide for yourself whether you feel ready, if you are unsure stick with the rules.
Concerning balance between characters, it is typical that characters are not on equal footing in all situations and it is good that way because a) it creates team play and opportunity for everyone to really shine and b) customization is an important reason to play RPGs. Taken to the extreme, however, like in your case, unbalanced parties will almost always create unfun gameplay, because a too powerful player can “monopolize” the game by resolving most challenges leaving no space for the other players, and because challenges will often be either trivial for the powerful player or problematically hard for the others. Therefore, all players need to adhere to the same rules to avoid this. It is important to note, that rolling for stats, even with the same rules for everyone can always create the same problem if someone rolls particularly high or low values. Many people like rolling scores and it is traditional but the potential of this happening needs to be considered when deciding how to determine scores. I can remember a case of this happening, I was a player in this case and it was D&D 3.5 but the same problems could appear in 5E: I dont’t remember the exact numbers but we rolled stats and I got lucky with I think no score under 10 and most over 14 and another player got rather low scores and it created feels-bad moments because my character was so strong in combat and the other one often felt useless in combat.
