Magico Q7 . . . wow

I stand partially corrected regarding the driver design. Stronger magnets will, indeed, increase the efficiency. But more than likely, it's just the woofers that make up that difference.
Tweeters that produce 94dB is easy, especially with neodymium. The same with midrange.
Woofers are little more difficult. More than likely, the original tweeters and midrange are now able to perform without being attenuated to match the woofers in the Q5.
This is a good thing, no matter how you look at it. Removing or reducing resistor value does help with dynamics.
Higher spl does usually equate with greater dynamics, all else being equal.
So I'll split the difference. I still think the tweets and midrange are the same. Bigger and badder woofer design now help in more ways than size alone.

As far as my musical tastes go, it is literally everything. Which is why none of these speakers find they way into my home on a permanent basis. It's difficult, but I require my speaker to perform well on all types of music.

Think ill jump on Q3 use my MAGICO MINI 2 for country home.How big a room do you need for Q3 or Q7 ? , Thanks.I do think the MAGICO MINI 2 is the best looking MAGICO sounds grest as well.

stereotaipei, could you provide to me at least two drivers of the similar size which eficiency(measured with 1w input, not 2.83v) varies more than 7db in 50-300hz range? I have measured more than 160 diferent drivers on identical conditions, so I think I know a litle bit what I am talking about.

Q5: 86dB
Q3: 90dB
Q7: 94dB
You are right, 8dB

I see we are discusing in diferent worlds. I am speaking about main part of the speaker- speaker drivers(woofers)and there is no posibility more than 6db diferencies in low FR on same size woofers. You are talking about speakers eficienty where sensitivity can vary from arround 80db to 95db (in some cases like horns vs electrostats 75- 113db) because not speaker driver(s) determine speaker sensitivity but speaker designer -enginier. he can reduce (by changing BSC etc. ) or tilt up ( by adding more paraleled drivers, changing drivers interaction etc)

so we are both right. btw In your place,I would take manufacturer specified sensitivity values with salt and grain- unfortunately there very rarely tells truth.