Omega Directive 24B Unslop v2.0 is a 24-billion-parameter merged language model built on the Mistral Small 24B architecture. Its creators combined two distinct finetunes using the Arcee Fusion method within LazyMergekit, weaving together ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0 and Doctor-Shotgun/MS3.2-24B-Magnum-Diamond into a single model. The "Unslop" designation signals reduced safety filtering and more direct response behavior, making this model purpose-built for scenarios where traditional guardrails would be overly restrictive. This design positions it as a strong candidate for creative writing, advanced role-play, and instruction-following tasks that benefit from unfiltered expression.

The model represents the latest evolution in the MS3.2 family, inheriting the combined strengths of its parent finetunes and optimizing for stable performance across extended context windows. It supports deployment in local environments via Hugging Face and Ollama, with multiple quantization options to accommodate varied hardware configurations. The merged architecture aims to deliver balanced, robust performance across complex language understanding and generation tasks, providing researchers and developers with a capable base for experimentation and domain-specific fine-tuning. Its architecture supports both bfloat16 and float16 precision for efficient inference, enabling practical deployment beyond cloud-based constraints.