Things are looking a lot better, and a lot of issues people are complaining about (no choices, city spam, armor, etc.) are really all just balance issues. That's a good thing. However, it's a little daunting, because in my opinion many progression curves are completely borked, which leads to a pacing experience that's like a toy boat in a hurricane. Here's the progression issues, and what might be able to be changed to fix it.

Champion progression

Until late game, champions completely dominate. Sometimes they still dominate at end game as well, but that's due to some unfair combos like fireball + evoker, not the base progression itself. The reason for this is very powerful abilities available very early in the champion's level curve, and very quick access to items and equipment.

Both of these are very easily solvable. Just move super powered stuff (charge, impulsive, evoker, high level spells, etc.) later. Paths are also very OP when you first get them, and it provides a huge jump in power for champions relative to everything else in the game. Maybe they should be toned down a bit, or players should be provided an option of "upgrading" their path to improve the bonuses every 4 levels or so (basic path of the mage, apprentice path of the mage, etc.)

The second issue is with super quick access to OP equipment. Because they can level fast enough to start taking out stronger monsters, champions get access to way better equipment way faster than units do. There needs to be a way to cap this. This _might_ be able to be done naturally by simply lowering the chances that equipment will be provided, or lowering the average "equipment level" dropped from quests or monsters. However you might want to look at something more artificial as well, such as a maximum level of equipment based off the number of turns played.

Army progression 

Armies suck early game. You get a free hero, and after a couple of levels that hero can kill many armies, so why even build low level armies except in some rare cases? This affects the city spam issue as well (as detailed in the next section).

Many threads have been created about this, but however you want to go about it, armies need to be much more powerful much more early. A group of spearmen should be a serious issue for any level 1 or level 2 champion without equipment, as should a group with shields. I understand you wanting to keep the very basic militia units weak, but the first tier of upgrades should be more meaningful, otherwise they will be worthless because of being over shadowed by champions (even before the champions start getting really good equipment).

Army levels are also a bit too powerful. Consider that each "level up" is potentially multiplied by 9, and most players might be surprised to find it's actually a lot better to level up a single good army unit than a hero.

City Quality vs. Quantity (city spam)

So I just learned that having more cities lowers growth. I should have noticed it right away, and not because the game goes out of its way to tell me. After realizing it I'm still going to city spam all over the place, because it's better. Why is it better? Because higher level cities aren't much better than lower level cities, so you might as well have a ton of lower level ones.

This is the easiest to fix. Make higher level stuff more costly to build and make higher level cities more effective. Right now buildings and materials have a much stronger effect than population. Make population matter more and the city spam issue will be mitigated, if not solved.

Also, providing armies as a production imperative earlier will make many cities less desirable, as well as limit the time that can be spent pumping out pioneers (assuming pioneers become reasonably costed, another issue that should be looked at changing). 

Spell progression

Almost all spells are coming from combat, and usually there is access to these very soon. This causes an issue where spell casters grow even quicker than other hero types, then plateau and stop any meaningful progression.

As mentioned before, moving spell levels to have certain level requirements will help this problem. Also, moving more spells into the magic tree will both make that tree more interesting, and provide a more reasonable spell progression.

I'm going to suggest something a bit more unorthodox as well. I like the idea the spell casters discover lost knowledge as they adventure, but abstracting it out to a level system is unintuitive and provides balance issues no matter how to play the numbers. Instead, I propose all spells past level 2 are learned through the magic tree (IE. research "Fire mage" spells and get access to all those spells), and that completing quests or exploring ruins provides a "magic tech" resource of some sort that can be used to speed up the research of these spells. Might fit a bit awkwardly into the current system, but could improve it quite a bit if it can be finessed in.