The problem is, you can’t just say “is Homey zigbee really this and that”, in general. It’s based on user experience, and the willingness to Reading (and understanding) The ‘Fine’ Manuals (zigbee is NOT plug and play, unless you’re lucky), and tuning wifi 2.4GHz signal channels (yours or close neighbour’s networks).
In general people only post complaints. Plus, they never tell if they’re using Thread as well, which can be quite a zigbee killer (read about Multi-PAN below);
So, say there’s 200 complaints, it’s very well possible 50,000 Homey zigbee networks are running fine.
No one knows for real, but we have to keep the reality ratio in mind imho.
Next,
The zigbee 3.0 standard is not a rigid standard like z-wave; Unfortunately manufacturers are allowed to add their own “flavour”. They mainly do that to make you buy their hubs/bridges as well, and ‘force’ you to stick to their brand.
Tuya (+ 100s of Tuya-whitelabels) and Aqara are notorious examples of “doing as they please”.
So, the less brands you mix, and avoiding Tuya and Aqara, certainly will have a positive impact on HOMEY zigbee.
The mentioned issues are almost non-existent with zigbee2mqtt, I can fully recommended it.
Most overlooked: this is no guaranteed way to create a stable zigbee network, there’s only best practice guides, like this one.
Thread & zigbee Multi-PAN (like Robert also mentioned):
This “invention” to share the hardware for both similar protocols turned out to be quite a bad idea.
However, Athom decided to just keep it like that @ Pro 2026 (and Pro mini?).
My advice: decide if you want to use zigbee or Thread, to avoid sluggishness, hickups and unresponsiveness.
Keep in mind this effect will be minimal with, say, 10 Matter-over-Thread and 10 zigbee devices, but it can become a pain in the ass when you extend the number of devices on either one protocol.
