"Is the uplift proposal available in the library?"

"I'm not actually sure! It was a relatively early-in-the-year project, though, and had to be purely technological, I'm sure I can dig up something better." 

(Kanimir is not actually familiar with the words "virus" or "pheromone" in context, and is trying very hard to Not Ask right now. Instead he is playing with one of the professor's aptitude puzzles.)

Kiri starts fidgeting with one too. "I think that was all the questions we had for you today."

Once the herd behavior squad (the... good guys?? The..... Control... Minders? The... they can workshop that) are all buddy-systemed to safe rooms, Kiri dives in to the aptitude tests. She is so curious what aptitudes they're testing and what kinds of magics use each of these faculties!

She isn't slow at doing things but performs worse under time pressure. She's creative when that means "whimsical" or "artistic" and less so when it means "think outside the box, but like, in the same way as the puzzle creator". She's weaker on underspecified problems - especially any that seem like they could be underspecified for reasons of poor puzzle design - than on exhaustively charted combinatorial explosions with fun exploits that pop up if you mix things weird. She avoids setting any puzzles on fire in frustration, but she does have to put down and come back to a couple more than once over the course of her puzzling. Her favorite is one that looks like a glorified dentist's waiting room fidget toy, where setting up background processes in certain orders with good internal feedback mechanisms adds up like liquid clockwork to any of a few dozen win conditions, but she's also pretty handy with a booklet of cryptanalysis problems and one with geometric constructions to choose feng-shui-esque layouts for notional magic objects. (She uses scratch paper for her solutions, so they can trade the puzzles around.)

She writes up all her results and opinions neatly to turn in to the magic teacher when everybody else is done too.

Erek is a robot and accordingly good at puzzles where they can be figured out by brute force of a million fiddly parameters. If he can do something at all, he can do it very fast. He's not uncreative but his creativity, if analyzed, would mostly look like he's turning up and down a "granularity" slider on the mix-and-match function. Perhaps incongruously, he's also strong on gradually incremental tasks where you build up or entrain something over many many steps, like bonsaiing a tree or training a dog.

His standout best puzzles - at least based on what the pack of benign students can guess about the background skill distribution - are a game that simulates domesticating tiny creature-esque motes by way of selectively harvesting them and putting all the puzzle objectives into the most efficient order over the course of their specializations; one that involves rendering geometric constructions including the ones that you can't do with just a straightedge and compass, though his come out very symmetrical and unadorned unless the constraints of the problem require otherwise; and multidimensional Sudoku-meets-nonograms-with-colors-and-shapes. The one he takes back to work on some more after everybody's had a first go at everything is the creature-mote one. On his first try he conservatively lowballed his expectations of how short he could get the generation times on his quick-dividing variant and he can get another bonus achievement out of the thing now that he's tried it once.

Wei Wuxian is smart in almost the exact opposite direction. He makes massive intuitive leaps, and even if you get him to sit down and try to explain the chain of reasoning he still ends up glossing over "obvious" steps that really aren't. He's very, very good at using the shape of the puzzle to deduce the creator's intentions, unless he instead gets his own idea for what to do with the puzzle and completely neglects any angle referring to what anyone else thinks. His approach to underspecified puzzles is to come up with a better one. 

Generally speaking if he and a puzzle have a disagreement it is the puzzle that experiences frustration. 

Edie is just a little bit less methodical than Kiri and Erek, but overwhelmingly more so than Wei Wuxian. Her particular strength is making connections--noticing when a puzzle has aspects that are similar to aspects from a previous puzzle she's already solved and using that to figure out the solution to the new puzzle. She has a fantastic working memory--nothing like Erek's, but enough that it's basically got to be part of her mutant power. 

Lucy doesn't make the bizarre leaps that Wei Wuxian does, but her educational background and set of techniques for problem-solving are more different from Edie's than Edie's are from Kiri's, despite the fact that both she and Edie were born on an Earth and Kiri wasn't. Her working memory is nothing special in terms of being able to hold onto independent facts at the same time, but the "facts" that she can hold onto are a lot more informationally dense than a human would be capable of--she can look at a scattered array of dozens of marbles and know how many there are as easily as a human would be able to if the number were instead four. 

Also, she cheats. A lot. If there's a piece of information she needs to hold onto that won't fit comfortably in her working memory, she creates a word for it in the Correspondence and moves on. She doesn't cheat in the sense of actually doing magic to the physical structure of the puzzles, but her spatial awareness makes that approximately moot.

Kanimir is very, very smart. He doesn't make Wei Wuxian's leaps of logic, but he's usually able to follow them. And he does make great intuitive leaps of his own, sometimes, they're just less buckwild and also he can articulate his chain of logic after the fact better. 

The kind of puzzle Kanimir does best at is when there are discrete components that you can use to build into any of a wide variety of configurations. 

Emily's smart, and she doesn't do badly on the math-related stuff, but where she excels is in more free-form, aesthetically driven areas.