Heroes 5.5 WarGame Edition - Tactical Edition.

The point is that it's not 5, but 4.
Because 2 are class-specific. One is purely class-specific, the other is a general class skill for the race.

I understand, there was a typo. 4 slots is not a problem.

There is a difference. One class allows you to enhance your native troops, while the second class allows you to gain additional troops with some new features. At the same time, of course, you can introduce perks in the general skills that will be available to one class and not to another. And all this within the framework of one race.
And the stats are, of course, also different.

So, I'm saying that significant differences are ensured even without additional entities.
Have you seen the scheme? The chances differ for 3 out of 4 stats, and 3 out of 6 non-racial skills (with a chance of 8% or higher) differ. There are hardly more differences than similarities.
Against this background, I specifically refused to introduce additional differences, such as class-unique perks or special class connections, because this is an excess that complicates the memorization of the skill wheel.
Personally, I can now remember the skill wheel for all classes even without practical gameplay, and yet, after many years of practical gameplay, I still haven't learned any of the wheels in the original game.

But why "cut it off"?
You can give magic to the Orc (surprisingly). And there are such tasty things in it as Seals and Weaknesses, for example. That is, in total, Diminishing Darkness (as a perk) + a couple of Darkness perks + Darkness itself with a Seal and Weakness can completely destroy the opponent's Darkness.

Magic won't technically work for the Orc.

Well, and as far as I remember, the Orc is still deprived in terms of Weakening Strike, because the effectiveness of Weakness on attack depends on the level of Darkness, and the Orc cannot take Darkness.

Weakness here gives -8/10/12/14 to attack, which is equivalent to -16/20/24/28 in the original, so it works perfectly even without the skill.

Added 59 minutes ago
P.S. I'm probably not very open to new ideas right now; I'm more interested in the practical implementation of what has already been conceived. For example:
- how to ensure that there is an opportunity for an ultimate expansion on the map - in 24466 (second city on day 2, 3-4 cities on day 4, 5-6 cities on day 6)
- how to ensure that it is possible to counter this with a rush on the first week
- how to ensure reconnaissance of the opponent's actions on the first week, including being able to detect the above-mentioned rush and expansion
- how to ensure the ability to defend against any rushes without artificial restrictions (like "restart on encounter" or "ban on red rush")
- how to ensure the division of troops between heroes
- how to ensure proper development of new captured cities as full-fledged bases
- how to ensure that the player can complete a turn in 3 minutes + only conduct 1 battle manually, while not feeling a lack of time
- how to ensure that the first clashes begin on the first week, but the game ends only on the 2nd-3rd
- how to ensure a series of interesting clashes instead of 1 "final battle" with the possibility of quickly replenishing competitive troops, going into an exchange, and not accumulating
- how to ensure maximum interest in development, given that the breakthrough itself takes a back seat (80% of battles are auto-battles)
etc., etc.
That is, purely practical questions, not of the type "what new thing should we come up with."

A class skill does not solve any of the pressing problems, neither general nor specific (like how to make a racial elf work), so I am not very enthusiastic about such ideas.