From the Wikipedia list, the most recently launched nuclear reactor appears to be a TOPAZ-I on Kosmos 1867, launched by the USSR in 1987. It's parked in a 800km orbit at 65º inclination and appears to be falling apart.
For Earth orbit applications, solar power mass efficiency (~100 W/kg) is much better than RTG (~2.8 W/kg) or reactors like TOPAZ-I (~15 W/kg). Even out to Mars or Ceres, solar still wins for mass. This means that almost all space missions rely on solar, driving R&D and bringing costs down.
For outer-planet missions, RTGs tend to win out even though they're less mass efficient than reactors, because they are simple and relatively cheap.
There are legitimate concerns about launching nuclear reactors on rockets that do, still, occasionally explode; no one wants U-235 scattered over thousands of square kilometers of Earth's surface, even the oceans, and even more so, no one wants to be the person who signed off on the risk assessment that let the launch go ahead, so for a bureaucrat, it's always safer to kick the risk assessment back for one more re-evaluation than to accept it.
For any given mission design, it's probably cheaper and simpler to reduce power requirements to the point that RTGs can handle it than to get a reactor in the design.