Prepiva’s free CELPIP mock test ends with all eight Speaking tasks, with the test’s preparation and speaking times, and the Speaking practice test runs them on their own. With a free account, each recording of your first full mock test and of your first Speaking practice is turned into a transcript by a speech-recognition model, and an AI language model rates the transcript on the four dimensions, giving each one a level from 0 to 12 with a reason, plus one strength and one next step.
- Timing counts: the rating also sees how fast you spoke, your long pauses and fillers, and how much of the time you used. An answer that fills less than about 40% of the speaking time is capped at level 4 for Content/Coherence and Task Fulfillment.
- Your Speaking level is the average of all the dimension levels across the eight tasks, rounded, with a tie rounded down.
Its limits: the rating reads a transcript, so it cannot hear your pronunciation, stress or intonation, which CELPIP raters judge under Listenability, and a speech-recognition mistake can change what it reads. It is one AI model, not a panel of trained raters, so treat the result as a practice estimate, not a CELPIP result.