Waveform models play a critical role in extracting valuable physical insights from gravitational wave data. As detectors become increasingly sensitive with each observing run, there's a growing demand for waveform models to increase their accuracy and computational efficiency. In this context, I will provide an overview of the ongoing efforts within the waveform modelling community to cope with these challenges. I will put specific emphasis on the efforts within the "phenomenological" waveform family to describe the intricate effects of subdominant modes, precession, and eccentricity.  I will end with a discussion of different techniques to accelerate parameter inference and reduce its computational cost.