Okay, sorry if I'm being stupid, but I can't for the life of me figure out why multi-classing casters would be useful. A caster with a martial, yes, but two casters seems worse than a single caster class. the main reason is because you loose higher leveled spells. Sure, maybe it makes you versatile, but that's why a wizard exists. Since you loose out on higher level spells, it just seems weaker to me. You can upcast spells, but why use level 3 burning hands when you could use fireball? I know that specific ones *sorlock cough cough* can be really strong, but won't they always fall behind at higher levels? 

Warlock dips are fun, you get a whole grab bag of different goodies depending on how much you put levels in, 1 gets you two invocations, some cantrips and a short rest spell slot. 2 gets you even more and access to agonising blast.

Delaying access to higher level spells is the main weakness of it, yes. But most of the spellcasting classes have unique features that can be used with other classes' abilities and spells, which can be really useful if exploited well.

One obvious example is that having levels in Warlock provides additional spell slots that recharge on a short rest and can be used to cast spells from other classes. In many cases they can also be used to power features from other classes that use spell slots to do things other than cast spells, such as Bardic Inspiration and Wild Shape, or converted into Sorcery Points. Sorcery Points themselves are also a big draw for multiclassing into Sorcerer, since Metamagic abilities can be used with any class's spells. 

Another option is to get armour proficiency to boost AC. One of my characters was an Artificer 1, Wizard X, getting spells a level late was a downside but made up for by higher AC, and the ability to cast cure wounds if the cleric went down.

Quote from Jegpeg >>

Another option is to get armour proficiency to boost AC. One of my characters was an Artificer 1, Wizard X, getting spells a level late was a downside but made up for by higher AC, and the ability to cast cure wounds if the cleric went down.

That's also a very good point. Another classic example of this is starting with a level of Paladin, thus getting proficiency with all weapons and armor, and then carrying on with another Charisma-based spellcaster class afterward. 

I think, also some subclasses can work well together. Like the classic storm sorcerer/tempest cleric where they have non-spell powers that can really pack a punch. 
Personally, I don’t like to multiclass, pretty much ever. But there’s some cases where it can really work. 

I’ll add, sometimes you might for character/story reasons. I don’t love those, always, as you can always rp someone as being really religious without taking levels in cleric, for example. But some people like the idea of it, even if it might be sub-optimal mechanically. 

I don't want to sound like a downer but if you start a campaign at level 1 it is extremely unlikely that you will progress to level 12. There are some class combinations that can shine at the lower/mid levels and there is no worry about high level spells because you're not going to get there anyway.

There is also the role play aspect where a player will chose a less optimal combination for the sake of story.

Quote from Wysperra >>

I don't want to sound like a downer but if you start a campaign at level 1 it is extremely unlikely that you will progress to level 12. There are some class combinations that can shine at the lower/mid levels and there is no worry about high level spells because you're not going to get there anyway.

There is also the role play aspect where a player will chose a less optimal combination for the sake of story.

 You do however need to worry about not being able to cast highER level spells. While you can enjoy several benefits of multiclassing you caster at (say) level five you are likely to miss not being able to cast things like fireball, hypnotic pattern, and spirit guardians.

What level you reach does depend on the campaign and group. I am in two campaigns one started over 3 years ago and the other nearly 4 having reached level 15 in one and 17 in the other. I also occasionally play on a Westmarch server where getting PCs to level 20 is quite possible.

The key to MCing casters is what level you do it at. It makes no sense to MC a caster before level 5 unless it is 1 level for armour proficiency. Otherwise you need to get to level 5 as fast as possible since most casters get a scalable character defining spell at 5th level. Most of the time, you are then free to MC and simply upcast your defining spell - be that fireball, conjure animals, or spirit guardians. You can usually keep up throughout tier 2 like that, but if you get to tier 3/4 you'll lag behind again.

Upcasts are generally poor returns even with a 3rd level spell compared to straight higher level spell options. Imo, the juice is very rarely worth the squeeze even for armor before you’re into tier 3- if your caster is constantly being attacked and you don’t already have decent armor you’re standing in the wrong places or your DM is aggressive enough about combat they’re likely to compensate for the armor. I’d maybe go for a 3 level dip if there was really good synergy in a feature, but generally you’ll get more powerful effects from higher level spells.

Quote from Wysperra >>

I don't want to sound like a downer but if you start a campaign at level 1 it is extremely unlikely that you will progress to level 12.

 Depends how the DM does level advancement; leveling can be fairly fast if using waypoints or in a combat heavy game, though of course probably half of all game die in six sessions or less.

Quote from Pantagruel666 >>

Quote from Wysperra >>

I don't want to sound like a downer but if you start a campaign at level 1 it is extremely unlikely that you will progress to level 12.

 Depends how the DM does level advancement; leveling can be fairly fast if using waypoints or in a combat heavy game, though of course probably half of all game die in six sessions or less.

 I agree with that but it also depends on the group, even if advanment is (relatively) slow. In my two campaigns, one was with some friends I have played with quite a bit on a west march server, the rest of the group had finished a campaign and one player decided to to drop out of the next one so they asked people on the server if they would like to join, I was selected because they knew I was reliable and my gamestyle match the rest of the group.  I entered the game knowing it would likely go to at least level 14 and was expected to 2 or 3 years at least..

The other campaign was a random group originally advertised on roll20, it is surprising that game is still going after nearly 4 years. 

Yes I understand. We've all been in campaign that last a long time and even get to higher levels. I've been two that legitimately went to 20 (from 1); a handful that went from 1 to 12 or 14; many that went from 1 to 5. These are campaigns that started and came to the campaign conclusion.

But overwhelmingly the number of campaigns that start then die before higher levels out-number the total of all of the above. Many many didn't even get past level 1. Several also don't finish - the players drop out or the GM goes afk for weeks, which leads to the players dropping out.

Campaigns that begin at 8+ give you a chance to plan out how much you want to sacrifice.

I like MCing wizards and bards. Wizards are obviously the best class, and when you combine that with a few levels of bard you can cast some healing spells too. There is nothing like casting ice storm one turn and mass healing word the next.

Meh, a 5 level investment is pretty high for any multiclass, but particularly full casters. If you want a bit of healing there’s feats, otherwise you can just go Bard if you want to diversify your picks without losing progression. Plus with a dip like this you’re splitting your casting stats.

Quote from AmberAundguard >>

I like MCing wizards and bards. Wizards are obviously the best class, and when you combine that with a few levels of bard you can cast some healing spells too. There is nothing like casting ice storm one turn and mass healing word the next.

There again a single class bard or tempest cleric can cast ice storm one turn and mass heal the next.

Quote from The_Ace_of_Rogues >>

Upcasts are generally poor returns even with a 3rd level spell compared to straight higher level spell options. Imo, the juice is very rarely worth the squeeze even for armor before you’re into tier 3- if your caster is constantly being attacked and you don’t already have decent armor you’re standing in the wrong places or your DM is aggressive enough about combat they’re likely to compensate for the armor. I’d maybe go for a 3 level dip if there was really good synergy in a feature, but generally you’ll get more powerful effects from higher level spells.

 IME level 3-level 5 spells are pretty even in power, and up casting can generally keep up. But for level 6+ spells you are absolutely right. 

Quote from Agilemind >>

Quote from The_Ace_of_Rogues >>

Upcasts are generally poor returns even with a 3rd level spell compared to straight higher level spell options. Imo, the juice is very rarely worth the squeeze even for armor before you’re into tier 3- if your caster is constantly being attacked and you don’t already have decent armor you’re standing in the wrong places or your DM is aggressive enough about combat they’re likely to compensate for the armor. I’d maybe go for a 3 level dip if there was really good synergy in a feature, but generally you’ll get more powerful effects from higher level spells.

 IME level 3-level 5 spells are pretty even in power, and up casting can generally keep up. But for level 6+ spells you are absolutely right. 

 
I‘ve crunched the numbers recently; it’s not a massive gap but the difference in averages can be notable. Plus in terms of secondary and other non damage effects it’s pretty much impossible to simulate with an upcast.

A 3 level dip delays but doesn’t prevent higher level spells. What it costs, if you go to L20, is your spell mastery ability, epic boon and signature spells. It also forces you to spend more experience for each spell casting class. After a 3 dip your L1 caster costs 2400 EXP more than going straight caster and L6 costs 34,000 more. This is a cost we often don’t really factor in to our calculations (or discount as being only 900 EXP). Between the delayed access to spells, the extra experience costs and possibly the tier 4 abilities the costs are generally greater than the benefits. Adding a caster class to a martial can make sense but the hidden costs are still fairly high.

One thing that no one else has mentioned when multiclassing casters is that it increases the number of spells you can have prepared. In the 2024 rules a level 16 single class full caster can prepare 18 spells not including automatically prepared spell. A 13/3 full/full caster character can prepare 23 spells. A 13/3 full/half caster character can prepare 21 spells. Warlocks follow the half caster paradigm for spell preparations. When you factor in automatic preparations through subclasses and class features this can vary between 4 to 11 extra preparations for the levels above.

That increase in the number of spell options at your disposal really makes a huge difference especially when you're primary class can change their preparations on a long rest like the wizard, cleric, and druid. Druid/rangers and cleric/paladin multiclasses even allow you to boost both of their DCs with the same magic items from Tasha's. With druids and rangers both using wisdom it's pretty nice to not worry about different DCs.

Edit: To push back on the dislike of upcasting spells there are some spells that upcast incredibly well. I'm currently playing a 13 sea druid/3 gloomstalker and hold person upcast to level 8 just ended an entire encounter against a bunch of bandit crime lords and deceivers. You can do something similar with confusion, Tasha's hideous laughter, and hold monster. Upcasting the right spells can have the same power as higher level spells.