The device on the left is a simple T-peice. It can be used for near DC operation. It can also be used at moderate frequencies (up to tens of megahertz, maybe a bit more) to produce a short (the shorter the better, usually the T-peice is attached directly to the equipment) branch from a transmission line to a high impedance receiver. The latter use is seen in 10BASE-2 Ethernet, CCTV, monitoring signals with oscilloscopes and probably many other applications. The advantage of such a set-up is it means you don't lose signal strength with each piece of equipment you hook up, the downside is that the stubs into the equipment can produce reflections which become more significant at higher frequencies.

The device on the right is a resistive splitter. Basically a T-peice with three resistors inside for impedance matching. Since this is impedance matched and only relies on resistors it can work anywhere from DC up to GHz frequencies and you can have long cables on any of the ports. The downside is it comes with a significant penalty in signal strength, the signal loss through the splitter (assuming all ports are correctly terminated) is 6dB.

Neither of these splitters provide "isolation", signals can travel from any port to any other port. Depending on your application that may be a problem or it may be irrelevant or even desirable.

---

There are two other types of splitter you should be aware of, likely they will look physically similar to the splitter on the right. Both are "power splitters", that is ideally they should result in a 3dB signal loss as the signal power is split equally.

One is transmission line based splitters, like those pictured in Bimpelrekkie's answer. These can be very efficient, but they only perform well over a narrow band. More complex shapes can widen the band but still there are severe limits on wideband performance.

The [first one pictured in Bimpelrekkie's answer](https://www.instockwireless.com/power_divider_pd1040_fab.htm) gets an impressively wide bandwidth for a transmission line splitter with about a factor of four between minimum and maximum specified frequencies.

The second one he pictures is much simpler and almost certainly has a much narrower bandwidth. Unfortunately it is sold by sellers who are clearly either ignorant of what they are selling or outright lying and claim it is suitable for "30-1000MHz" which is clearly bullshit.

The final type of splitter is a transformer based splitter. These can give good performance over a wide band, but they don't get down to DC and they tend to be lossier than transmission-line based designs at microwave frequencies, for example [here is one from mini-circuits](https://www.minicircuits.com/WebStore/dashboard.html?model=ZAPD-2-252-N%2B) which is specified over the range 5Mhz to 2.5GHz, though the loss gets noticablly higher towards the upper end of that range.