In this letter, we compare the capacity of a massive multiple-input multiple-output (MIMO) system using a low-resolution analog-to-digital converter (ADC) and a linear detector against a conventional MIMO system with higher order modulation and near maximum likelihood (ML) detection. We show that in the low-signal-to-noise ratio (SNR) regime, the quantized massive MIMO system can outperform the conventional large MIMO system; however, for high SNR, the conventional MIMO system with a near ML detector can outperform the extreme 1-bit quantized massive MIMO system. An analytical framework that derives the achievable rate of a linear minimum mean-squared error (MMSE)-based detector in a massive MIMO configuration, with the assumptions that the front-end is limited to a low-resolution ADC and channel estimation is imperfect, is presented.