CsoundQt offers different ways to run/render Csound code.
When the Run button is pushed (or Control > Run Csound), the current code is executed in real time. This means, the audio output is written to the soundcard (dac = digital-to-analog-converter), not to an audio file. You have any influence to the performance your code realizes, for instance by widgets or MIDI.
When the Run in Term (= Run in Terminal) button is pushed, the code is also executed in real time, but as a subprocess in a Terminal. The main difference to 1) is, that any CsoundQt specific code will not work. In particular, widgets will have no effect at all.
When the Render button is pushed, the code is rendered and the result is written to an audio file. You will hear nothing in real-time. The file name is taken from the setting in Configure > Run > File (offline render) > Output Filename (when checked).
When the Record button is pushed, the code in executed in the same way as in 1), but additionally the real-time performance is recorded to an audio file. The file will be written to the same directory as the csd file is in, and will carry a similar name. It is also possible to toggle record on/of several times during the performance. CsoundQt will write seperate files for each time, record was on.