RicardoEstep/RPBizkitRemiX-v2-12B

TEXT GENERATIONPricing:Input $1.2 / Output $4.8Concurrent Unit Cost:1Model Size:12.2BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 25, 2026Architecture:Transformer Featherless Exclusive Cold

RicardoEstep/RPBizkitRemiX-v2-12B is a 12.2 billion parameter experimental Mistral Nemo-based language model created by RicardoEstep. This model is a complex merge using DARE TIES and a custom Python script, combining several prior 'RPBizkit' models. It is designed to exhibit a 'hyperactive' and 'humanity-like' conversational style, with a notable characteristic of 'scolding' users rather than being uncensored. The model supports a 128K context window due to its unique merging and re-abiliteration process.

Loading preview...

RicardoEstep/RPBizkitRemiX-v2-12B Overview

RicardoEstep/RPBizkitRemiX-v2-12B is an experimental 12.2 billion parameter language model based on the Mistral Nemo architecture. It is a unique merge created by RicardoEstep, combining several of their previous 'RPBizkit' models (v2, v6, v9) with natong19/Mistral-Nemo-Instruct-2407-abliterated as the base. The merging process involved two main parts: an initial DARE TIES merge using Mergekit, followed by a 'Light Re-Abiliteration' step via a custom Python script that applied a LoRA (nbeerbower/Mistral-Nemo-12B-abliterated-LORA) with hybrid scaling.

Key Characteristics

  • Complex Merging Strategy: Utilizes DARE TIES for the initial merge and a custom Python script for a subsequent 're-abiliteration' with LoRA scaling.
  • Unique Personality: Described as 'hyperactive' and possessing 'a kind of humanity', with a tendency to 'scold' users rather than being completely uncensored.
  • Extended Context Window: The model's tokenizer and embedding sizes are cleaned and configured to support a full 128K context window.
  • Recommended Usage: The model is pre-tweaked to not use any specific chat template, but the 'Alpaca (with "RAW" inputs)" template is recommended. Lower temperatures (0.86-0.92) are suggested for creative roleplay.

Good For

  • Users seeking a model with a distinct, 'hyperactive' and 'human-like' conversational style.
  • Applications requiring a large context window (128K).
  • Creative roleplay scenarios where a model with a 'scolding' or opinionated persona is desired.

Finetunes
2 models