5e Wizard is the best Wizard of any edition, and I like the Wizard in every edition. The spell mechanics with spontaneous slots is excellent. The only better mechanic is the Short Rest spell points of the LaserLlama Psion. The Short Rest spell points system might make its way into the 2024 Dungeon Master Guide. The Wizard class might benefit from more thematic focus, such as only mastering two or three spell schools, where the schools are more like thematic domains, and well thought out. Not a fan of vancian casting, but thats the way it was in earlier editions. Glad 4e and 5e ended those mechanics. The 5e Wizard is better than the 4e one, because the 5e spell system is more ad hoc for freer spell selection (rather than AEDU), yet retains the 4e goal of balance.
5e Bard forever! The mythologically accurate full caster Celtic Bard is the only way to do the class properly. Additionally, the 5e bardic themes and spells are relevant for many Nordic (Norse, Sámi, Finnish) concepts, and other shamanic concepts from various other cultures around the world. The D&D tradition of Bard has always been an awkward mishmash of Arcane, Primal, and Divine. The Bard should be a Psionic class: the magic of the mind of an artist. This personal mindful source of magic power is likewise relevant for the Nordic shamanics. It is very important for the Bard to function properly WITHOUT a musical instrument. In this case, the Verbal component can function as the spell focus. From shaman to dancer to commander to the Celtic poet Bard itself, too many Bard concepts have nothing to do with a musical instrument.
I love the 5e Paladin. It is a muscle mage, that is so versatile for building character concepts from Tolkien Gandalf to Norse Thor, plus Jedi. I would like the player to choose which damage type the default Smite inflicts. Radiant is great for Gandalf, but Lightning-Thunder for Thor, and Force for Jedi. Etcetera with other kinds of principled tough mage concepts. In any case, I love the 5e version of Paladin.
Druid is a great concept, but its mechanics all over the place making hard to synergize concept and mechanics. I like the 5e Druid narrative. The Primal source is all about the Material Plane, and actually has little or no interest in the Feywild or Elemental Planes. It is the Fey and Elementals who sometimes take an interest in the Material Plane.
Warlock has the best mechanics in 5e. Its solid range of balanced Short Rest casting, always-on Invocation features, and access to highest tiers spells, plus melee competence, are a virtuoso of design. I want to see 5e Psion, Swordmage, even Warlord using the Warlock chassis, albeit with Short Rest spell points in place of Short Rest slots.
Barbarian. 5e is fun and leans into the Primal source. Its Berserker should be more shamanic, and animalistic, even with Wildshape capability. But as a whole, the class dabbles in this.
Fighter. Should split between heavy infantry heavy armor Knight and light infantry highly mobile Skirmisher. Skirmisher keeps the name "Fighter", with a nod to fight sports. 4e is ok, splitting up the Martial source into Fighter, Rogue, Warlord, and Ranger. Here the noncaster Ranger handles the Skirmisher, then Fighter handles Knight. Rogue is like special ops. By far my favorite is Warlord.
Rogue is fun. Probably the 5e Rogue is my favorite version. It holds its own, balances with other classes, can handle different concepts, and is fun to play.
The Cleric of every edition is too presumptive about what the setting must be − too fixated on polytheism. 3e is a bit more officially customizable, and 5e is loosening up, such as with Xanathars. I hope 5e 2024 will just once and for all let the player decide whatever "sacred" concept is appropriate for the character concept, and for the player and the DM to figure out a sacred community in whatever setting. Because of the meaning of the word "cleric" and "clergy", the choice of a specific sacred community and the official position the character has within this community, is the most salient aspect of the class concept. Mechanically, 5e is decent, by relying on "Domains" for thematic spells and how gishy.
Monk. 5e is ok. The base class needs to transfer much content to a specific subclass. Then the base has less design space and the subclass has alot more design space. Then the subclass can open up many different Monk concepts. Perhaps even call the class Athlete with the Monk as one of its subclasses.
Ranger in the sense of competence in the wilderness seems best as prestige subclass that any class can take. Then the nonmagical Ranger has the Fighter or Rogue as the base class. A Druid base yields a wilderness caster. And other concepts are possible as well.
The 5e Artificer disapointed me when it went partial caster. The 4e Eberron Artificer is best, an Arcane healer and magic item master, who can function as a full caster. I would love it if the 5e 2024 Artificer used the Warlock chassis instead.
The 3e Sorcerer was hero for putting a crack in the "vancian" mechanics spellcaster. But has done little since. The critique of the underpowered 3e Sorcerer versus the overpowered 3e Wizard, resulted in the creation of a fantastic 3e Psion and 3e Warlock. But the Sorcerer found little or no improvement. When 4e and 5e ended vancian casting, the purpose of the Sorcerer class remains in question. I find the fiddly mechanics of the 5e Sorcerer to be unpleasant. Meanwhile its supposed innate magic is absurdly sabotaged by its need to cast spells using a material component.