Degraded speech intelligibility due to dysarthrias can result from changes inherent to diseases such as Amyotrophic Lateral Sclerosis (ALS). Tracking this degradation longitudinally can highlight treatment effects. Various approaches have been developed to automatically measure speech intelligibility, often focusing on output from automatic speech recognition systems. In this paper, we extract phonetic outputs from the wav2vec 2.0 model and compare them to expected phonetic outputs across 12 languages to derive Articulatory Precision (ArtP). The strongest correlations between ArtP and Azure Pronunciation Assessment were 0.93 for English, 0.85 in German, and 0.66 for Swedish. We additionally find that ArtP has a correlation of 0.77 with the ALSFRS-R speech subscore and is sensitive to perceptual measures of speech intelligibility. The measure shows promise as a reliable and clinically relevant tool that can be used in multiple languages and disorders to assess speech intelligibility.