Amazing Writing Models: Cydonia-24B-v4.3 or Magidonia-24B-v4.3

I’d love to see the Cydonia-24B-v4.3 (based on Mistral Small 24B) given a try for Venice platform. I’ve been running this model locally, and the experience is distinct from anything else I’ve tried so far. It is a dedicated Storytelling/RP model that sacrifices coding/math capabilities to go "all in" on writing—and it shows.

What are these models about?

Based on Mistral models, Cydonia (Small 3.2) and Magidonia (Magistral Small 3.2) are RP/Storytelling focused LLM Model that really put it’s entire focus on Writing at the expense of other purposes such as problem solving or coding, which are set aside. 

What’s so special about these models? They’re small and have little parameters.
It’s that special emphasis on writing that makes it so impressive.

• Character Autonomy/Not user-centric: A pattern I noticed amongst other LLM models is that they seem to attribute the protagonist syndrome to the user. This model seems to steer away from that. If the LLM character is “popular”, for example, the conversations might be interrupted by other characters around while in the open.

• Large memory : 131K context window. I tested it up to 16K context and it continued to remember things from the 7K mark. 

• Efficient: Self explanatory, small.

It’s hard to explain without trying it, but I think it’s worth considering of exploring this model for Venice.