Heterogeneity of tinnitus imposes a challenge for its treatment. Identifying tinnitus subtypes might help to establish individualized diagnosis and therapies. The minimum masking level (MML) is a clinical tool defined as the minimum intensity of a masking sound required to cover tinnitus. Understanding the differences among masking patterns in patients could facilitate the task of subtyping tinnitus. Here, we studied the variability of hearing thresholds and MMLs among patients with tinnitus to identify tinnitus subgroups. A population of 366 consecutive patients from a specialized tinnitus clinic were included in the analysis. Hearing thresholds and MMLs were determined for octave frequencies from 0.25 to 8kHz, as well as for 3 and 6kHz. Subjects were divided into two groups according to whether their tinnitus was maskable (M, 329 subjects) or non-maskable (NM, 37 subjects). Hearing thresholds and tinnitus loudness did not differ significantly between both groups. The dimensionality of the data was reduced by means of principal component analysis (PCA), and the largest resulting components were used for clustering the data. The cluster analysis resulted in five clusters with differences in tinnitus pitch, lateralization, hearing thresholds and MML, as well as on age and gender. Clusters differed in contours of hearing thresholds and MML, describing patterns of low or high thresholds in combination with low or high MML. The clustering solution presented a low silhouette value (0.45), implying that the clustering is weak and could be artificial. The analysis pointed out the diversity across tinnitus patients. Our results suggest that there might be a continuum of patients' characteristics rather than discrete subgroups.