BeeChat v2 released: Bigger model - better performance!

We recently published an improved version of the initial BeeChat model, resulting in even better performance! In detail we switched from the default llama 3.3 model variant (Q4_K_M) to a variant with more precision (Q6_K).