RicardoEstep/RPBizkit-v10-12B

TEXT GENERATIONPricing:Input $0.87 / Cached $0.2 / Output $0.99Concurrent Unit Cost:1Model Size:12BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 28, 2026Architecture:Transformer Featherless Exclusive Cold

RicardoEstep/RPBizkit-v10-12B is a 12 billion parameter experimental Mistral Nemo mix model created by RicardoEstep, utilizing a Model Stock merge with Mergekit. This model is specifically fine-tuned for 'dark RP' (roleplay) scenarios, combining pruned weights from various base models to achieve stable creativity. It features a clean tokenizer and embedding sizes supporting a 128K context window, making it suitable for extended roleplaying interactions.

Loading preview...

RicardoEstep/RPBizkit-v10-12B Overview

RicardoEstep/RPBizkit-v10-12B is a 12 billion parameter experimental language model, a unique "Mistral Nemo" mix created by RicardoEstep. It leverages a sophisticated "Model Stock merge with Mergekit" technique, combining several base models specifically tuned for "dark RP" (roleplay) to achieve a balance of stability and creativity. This model is designed to provide a high-quality, stable, and creative experience for roleplaying applications.

Key Capabilities

  • Specialized Roleplay (RP) Focus: Built from a blend of "RP Uncensored" models, it is specifically optimized for "dark RP" scenarios, aiming for "pure stable creativity."
  • Advanced Merging Technique: Utilizes "Model Stock merge with Mergekit" to combine pruned weights from various base models, ensuring a refined and effective blend.
  • Extended Context Window: Features a clean tokenizer and embedding sizes (131072) that are "supposed to" support a full 128K Context Size, enabling long and complex roleplay interactions.
  • Optimized for Stability and Creativity: Addresses previous versions' issues by providing a model with enhanced stability and creativity, free from common merging problems like the "fix_mistral_regex" patch.

Good for

  • Immersive Roleplaying: Ideal for users seeking a model capable of generating detailed, creative, and stable responses in "dark RP" contexts.
  • Long-form Conversational AI: Its large context window makes it suitable for maintaining coherence and memory over extended dialogues and narratives.
  • Experimental Model Merging Enthusiasts: Offers an example of a complex model merge using specific techniques and a curated selection of base models.

Finetunes
2 models