If you really want to know the temperature, use a thermocouple probe. The various temperature guns depend on assumption of emissivity of the surface you are measuring. This can be quite accurate, or wildly wrong for a variety of reasons, many not obvious at first glance. With that as a background, I find that the cheap guns ($20-$40) work well enough to not be the dominant error source. As with any temperature measurement system you will need to take the time to characterize it. Figure out how large the measuring spot is and how accurate the red laser indicating spot is, and get a sense of how reflections, surface roughness and surface coatings can change temperature indications.