Based on the video [1] it seems like a real time text translation. Its doesn't do real time audio translation, which is what the title of article implies.
It’s transcribing your audio and translating it into a subtitle. That seems what would be expected, unless you thought it would synthesize a voice in Mandarin on the other end? That would probably not even be desirable, just as subtitled movies are better than dubbed movies.
Eh, I thought of text rather than audio when I read the title.
I also thought it was real time audio translation, as google meet has speech to text already for a while.