There are two basic types of electronic temperature sensors: thermocouples and resistance sensors (primarily RTDs and thermistors.)
Thermocouples have much less mass in the sensor itself, and can be grounded - meaning the sensor actually touches the inside wall of its protective sheath, giving a much quicker transfer of heat energy - for a non-transmitting sensor output. Ungrounded thermocouples, in which the sensor is isolated from the protective sheath, are slower than grounded thermocouples, but still approximately three times faster than a resistance sensors if the overall prove construction mass is similar. No resistance sensor is grounded to an electrically conductive sheath, as that would ground out the loop of current flowing through the sensor.
Thermocouples are generally about three times faster responding than a resistance sensor and response times are generally stated in a couple of ways. Some manufacturers will give two response numbers - for instance, a 99% change going from whatever ambient starting point the probe is in, to being plunged into boiling water. The second number is for a true 99% change from ambient to full process temperature being measured. 100% is generally not stated as the final tiny fractions of a degree change are difficult to pin down.
Resistance sensors, primarily RTDs, are more difficult to determine, so a time constant is used to be able to compare sensors against each other to determine their suitability for a particular application. Time constants are values used to indicate the time required to reach a 63.2% change from a starting point to that percentage of change in the step temperature. This time constact will give a response time relationship between sensors to indicate the best selection for the application. If a total response time to 99% is needed, a good rule of thumb is a total of five times the time constant for the particular probe.