Note: question changed since I wrote this answer.
Two classic and highly effective fractal antennas are the log periodic antenna and the many variations (planar, conical, cylindrical, etc.) of the log spiral antenna. The main characteristic of both of these antenna families is extremely broad bandwidth, which can be attributed to the fractal self similarity at multiple scales.
Most other fractal antennas are either less well known or are less effective than an isotropic antenna or the fat wire equivalent antenna, or are extremely specialized. (If anyone knows of others, they should add another answer here!)
Around 2000, when idea of the fractal antenna was first published, there was a lot of enthusiasm, but most papers didn't include any studies of the effectiveness of such antennas. In the following years, a number of additional papers were published, finding that all the new fractal antennas they actually characterized either had lower gain or lower efficiency (or both) than similar antennas that just used fat wires instead of fractal geometry. Fractal antennas did inspire some innovations in antenna theory around the nature of what increases the bandwidth of an antenna.
In a brief search of the literature, I see a lot of fractal antennas, but the papers I found either did not mention antenna effectiveness at all, or admit poor performance. I have heard of rumors of proprietary or trade secret fractal antennas in commercial use, but until something is published about them, they are just rumors. There are two that work, so it is possible there are others.