Since some people expressed concern in comments as well as DMs I just want to reiterate that i'm not going to switch towards anime/japanese style.
I'll be continuing using the current model until I can get the new one to look right. The pics I posted were to share my excitement at the conceptual understanding it showed. You can expect to see requests looking same as before while I experiment with the new model in my free time.
So far I've tried a more flexible version of the model and tested out about 540 artist prompts. Some results look like promising starting points to recreate my current style. Making a data set of my own work and training it on that is also an option, as is a two step process using both models sequentially.