The current study examined the relationship between overall second language (L2) proficiency and utterance fluency measures for several L2s in order to determine whether utterance measures can be used to predict L2 proficiency. The study measured the speech rate, number of hesitations, number and length of pauses, number and length of runs, and number of false starts using excerpts from 126 ACTFL Oral Proficiency Interviews (OPIs) spoken by 86 participants. Forty of the participants provided pre- and post-OPI speech samples, which also allowed examination of changes over time. All 86 participants were native English speakers who spoke L2 French, German, Japanese, Arabic, or Russian. They ranged in proficiency from Novice Mid to Superior. Results suggested that some L2 utterance fluency measures correlated significantly with overall L2 proficiency for all L2s, but data also revealed some differences across L2s. These differences hinged partly on the L2's relative difficulty for native English speakers. Results suggested that it might be feasible to use specific fluency measures to estimate proficiency, in particular at higher levels, but that fine-grained sublevel estimates would not be recommended, in particular at the Novice and Intermediate levels.