I have the Modi 2 multibit, and admittedly haven't heard the standard/ uber versions, so I can't (and won't) be trying to compare them. However, I can at least try to explain the difference between the standard (and uber), and the multibit and why it is "better" (quotes because it is subjective and based on personal preference like so many other things in the audio world).
DACs come with a specific number of "steps" that they can break a digital signal up in to. This will likely vary depending on the DAC that you are looking at, but for these purposes I am going to state that 16 is the "standard" (may be completely incorrect). That would mean that a digital signal sent from your source has 16 bits of precision. Typically speaking the more precision a value has the better it can be represented... but that is a simplification because it really depends on the type of number, etc. For instance a floating point number (number capable of storing a decimal value instead of just an integer) can have varying mantissas associated with them to give different values to the power section or the integer section of the number at hand, while an integer can only count by whole numbers (so 1, 2, 3, etc.). Either way, the more bits a digital number has access to the more values can be represented by it.
For example, a 16-bit digital value can store 2^16 different values ranging between 32,768 and -32,767 (0 is counted as a value, so positive or negative has 1 fewer possible values than the other... I can't remember which at the moment, so I chose negative arbitrarily. It is probably wrong, so if you really care I suggest googling it). This gives you a total of 65536 "steps" that can be represented in the sound wave from the digital signal. This also means that when you are converting the signal to analogue there will be only so many different values that can be represented.
Now, when looking at the multibit, it supports 24 bits of resolution. That means all of the sudden instead of having 65536 total values that the digital number can represent to the DAC, it can represent 2^24 total values (16777216) which is significantly more than a 16-bit value can represent. This means that the analogue graph _can_ (not necessarily _will_) have more "stairs" in it and be a smoother transition since each of the steps is smaller.
Basically what it comes down to is the more values your DAC is capable of converting from digital to analogue, the smoother the sound wave your DAC is able to output to your amp and subsequently your headphones/ speakers. Sorry, that last sentence sounds horrible, but I also haven't come up with a way I like better, so it stuck.
One other thing to keep in mind is that depending on the audio files you have access to and how they were recorded, they may have more or fewer bits available than your DAC can handle, so it is possible that you won't notice a difference because your music doesn't actually have a difference encoded within it. Likewise, having a 32-bit or 64-bit DAC would be even better than a 24-bit DAC, but most music has nowhere near that many bits of resolution at this point, so it would be a significant waste to get a DAC with the resolution since it would never be used (as of now).
I am not sure how well I stated that, but hopefully it gives you some insights.