Forty marks, thirty-seven questions, six parts, about fifty minutes. In the real exam every recording is played twice and there is a pause to read the questions first. Both of those are reproduced here: each extract gives you two plays, and the counter tells you what is left.
Read the questions before you press play. Half of listening is knowing what you are listening for. The other half is not panicking when a word goes past — keep going, because the answer to the next question is already on its way.
The French here is spoken by your computer’s built-in voice, not by actors in a studio. That makes it cleaner than the real thing: no background noise, no overlapping speakers, no regional accents, no one trailing off at the end of a sentence. It trains the mechanism — hearing numbers, holding a sentence in your head, resisting the distractor — but it does not train you for the messiness of a real recording.
One more difference: with a single voice, the dialogues here announce each speaker by name so you can tell them apart. A real exam recording uses different voices instead and never says the names.
For the messy end of listening — liaison, silent endings, words running together — use the Dictation Lab, which is built exactly for that.