I’m using the TheBloke/U-Amethyst-20B-GGUF and of all the popular models between 13b and 33b i found it’s the sweet spot. Not many regenerations needed and very good for roleplaying, not overdoing the storytelling and holds up the character card really well.
If you download GGUF models from “thebloke” you can read on the models card page how much RAM is required for the specific model without offloading to the GPU.
I have included a screenshot as an example of a 13b model.
https://preview.redd.it/3bv297j8c92c1.jpeg?width=1426&format=pjpg&auto=webp&s=94fac2937e8a2e0f6b3886d42401b0b50b0010b3