As mentioned several times in this thread before:
The color camera will only collect red light on a quarter of the sensor size due to the Bayer Matrix which is specifically critical for the Ha data.
For O3, the blue quarter + the two green quarters (thus 75%) of the sensor can be used, so the difference between mono and color camera for O3 is much less compared to Ha.
One should also note that mono sensors often have higher quantum efficiency compared to color sensors. All in all, I would prefer mono if you can.
To bring some real data on the table: Below, you will find two stacks of Sh2-129 using the same telescope and two different cameras utilizing the mono and color version of the Sony IMX571 sensor. Both stacks are taken from the same Bortle 8 balcony, same time of the year. Dualband from 2022 and mono from 2023.
Upper panel: 11 hours of mono data using an Antlia 4nm Ha-filter (Processing: Auto-strech + graxpert)
Lower panel: 16 hours of dualband data using an IDAS NBZ dual band filter (roughly 6nm for Ha), where only the red channel has been extracted (Processing: Auto-strech + graxpert).
As you can see, there is a lot more detail in the mono, and the noise is much less also.
I did not have meaningful data to compare also O3, but the difference is much less striking there.
11hr_mono.jpg
📷 16hr_dualband.jpg
-Martin