The problem with identical clones is they lack diversity. One disease will wipe them all out, there won't be small people able to wriggle through cracks and large people able to apply huge brute strength or tall people able to see further.

If you standardize everything into conformity then you reduce the range of different tasks your teams can perform. Instead you want a mix of different physical types across the range and then let each take the role they are best suited for.

Your methods are helpless against mutation.

After the initial injection of DNA into the egg, one can no longer regulate their biological development outside of nourishment and hormone treatment. This allows possibilities of mutation during cell growth.

A particularly dominant mutation could spread throughout the fetus/embryo and result in a soldier with their own quirks. Every now and then one is born with different eye or hair colours, a slightly greater intellect or, coincidentally, an unusual height.

The mutations should not be too radical 'by default', but they could be amplified as a side-effect of the chemical cocktail used to accelerate growth.

Alternatively, mutation inhibitors may have been invented to reduce the frequency or severity of mutations. These chemicals may be too expensive for extensive use and are reserved for an elite force (an interesting concept).

It may simply not be worth ejecting all non-conforming soldiers due to the cost of fertilising an egg/monitoring it constantly/feeding it until it is born. Therefore the soldiers will have to adapt and be trained for their standard equipment.

I hope this provides some inspiration.

Edit:

Having soldiers of various heights could aid remaining unsuspicious when executing covert operations, since otherwise a soldier of exactly e.g. 1.836m in height would be suspicious and taken aside for inspection at border controls.

In large groups, having soldiers of various heights would aid visibility of a singular position, similar to stages.

Varying heights could aid in combat specialisation, with lower centers of gravity allowing for more athletic maneuvers while taller soldiers could perhaps have more strength or leverage.

Short soldiers would take less space, less food, less material for uniforms and less weight among other things, making them ideal for mechanical deployment, sieges or other forms of long-term semi-isolation.

Taller beings would be better suited as look-outs or for transport purposes, being able to march quicker, wade through deep marshes and step/climb over obstacles more efficiently, but having a high center of gravity would not necessarily be advantageous as far as I can tell.

Different models for different jobs

The ideal infantryman looks nothing like the ideal pilot. The ideal pilot doesn’t necessarily look much like the ideal tank operator. The ideal tank operator certainly doesn’t look much like the ideal medic. An army is full of people who play different roles, and there’s no reason to make them all look the same.

Like father, like son

If you want a fun excuse for this, say that each job role has a ‘blueprint’, based on the most outstanding representative of that role at the time when this super soldier was begun. All the medics are the perfected ‘sons’ of a highly decorated combat medic, etc. This gives you diverse looking troops who are incredibly proud of their lineage.

Unification gives great economical bonus, thus specialization is about great efficiency.

Let take a look back: two centuries ago, grenadiers used to be tallest soldiers as they had longer arm and could do bigger swing movement while throwing a grenade (resulting a longer throw distance).
Contrary, hussar cavalry was about light body weight to improve horse attack speed (naturally leading to low-stature dexterous recruits).

So, your war economy model is way more complex than checkers vs. chess dilemma and should include to-be-born cost + manufacture associated equipment as well as war field efficiency metrics.

Something about the manufacturing process produces different results. Maybe the farther you are from the center of the incubator the taller you are. Or maybe there's some sort of layered, onion-like ingredient and the "strips" closer the center are more potent.

Anyway, there is this unavoidable consequence that each batch produces an array of sizes.

This has another advantage in that it enforces societal roles. Pilot's are more important because very few are produced per batch.

Chip Binning Anecdote

Interestingly enough, this type of manufacturing already exists! Have you ever wondered how Intel makes so many different models of processors? They actually only make a few (i3, i5, i7, etc.). All the variants, like i7 8700k, happen because the manufacturing process does not produce exact results. One wafer (batch) produces CPU's of all different specs. Here is an article that explains it (scroll down to step 10).

Just like chip binning, your manufacturing process could produce people in different sizes all in one go. These different sized people could then be "binned" into their proper positions.

In the past, supersoldiers were produced that way but it turned out failure to conform with the strict requirements lead to many being abandoned part way through production. In contrast, the requirements were found to be far too strict and with advancements in supersoldier manufacturing processes many previously discarded soldiers were found to be perfectly capable of less demanding tasks. Most supersoldiers are now binned to maximize yield.

The best supersoldier are the ones who win wars. Unfortunately, that doesn't tell us exactly which characteristics to give our supersoldiers.

A possible solution to this issue is to use a method similar to natural selection: To produce new supersoldiers, randomly select two existing supersoldiers, mix their DNA, apply a slight mutation and grow the new, and hopefully better supersoldier from the result. The random selection should be biaised toward the fittest soldiers according to some metric of your choice.

That way, the soldiers become gradually more efficient at war, which will ensure the dominance of that army for decades to come.

Using this method would mean that the army is made of genetically diverse soldiers, which implies varied sizes.

There are a few problems with this method (completely random DNA mutation produces a lot of waste, random DNA mixing may produce suboptimal results, etc...), but they can be fixed with sufficiently advanced knowledge and technology, but that should not be a problem for a supersoldier-producing civilisation.