Generally speaking on A2A COPs are a bit higher than A2W (because you have one less heat exchanger to go through and refrigerant tempatures can be lower too), and they are a little less temperature sensitive.
On my Daikin units, the data sheet list the COP as 4.2 @ -5 C OAT (21 C inside target). That's at max output too, at minimum modulation the COP is 4.75 @ -5 C. There's some caveats because they don't count defrosts in that so real world is lower, but we achieved about a ~4.1 COP is the recent very cold weather. Coldest day with a low of -7 had 22.5 kWh input power, compared to about 110 kWh on our gas boiler.
There's some other multi split catches too, because on some units efficiency drops significantly if you only have one unit running. On my Daikin 4 port unit, if you have one unit running compared to all 4, then COP drops from ~5.3 to ~3.2. You generally want to avoid this on affected units. That would bring me to:
I don't know the exact model of outdoor unit you have, but the MHI 8 kW (SCM80ZM) unit I found as an example at minimum modulation, running 1 5 kW indoor unit, has a minimum heat output of 1.1 kW, with a minimum input power of 390 W. Thats a COP of 2.82. Pretty bad. The same outdoor unit running 3 indoor units (A 5 kW, 2.5 kW and 2.0 kW) all at the same time has a minimum heat output of 1.6 kW, and a minimum input power of 370 W, now with a COP of 4.32. AKA it got 50% more efficient. (Source:
https://www.mhi.com/group/maco/products/downloads/pdf/sr_series_2022.pdf / Page 57)
Obviously if you only need heat in one room then you're pretty stuck, but it is worth bearing in mind that it's usually a serious COP hit to do that so best avoided if possible. Better to heat all of them together and turn it off once they all get up to temp. A2A has much, much lower cycling losses than A2W does because there isn't 100+ kg of water mass to deal with.