Torero96Dev/Cupid-Qwen3-4B-v0.2

Hugging Face
TEXT GENERATIONConcurrent Unit Cost:1Model Size:4BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Nov 15, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Warm

Cupid-Qwen3-4B-v0.2 by Torero96Dev is a 4 billion parameter LoRA fine-tune of the Qwen3 architecture, specifically optimized for non-reasoning, uncensored roleplay. This model aims to provide a fast and fluid roleplay experience, diverging from larger, reasoning-focused models. It features an updated and expanded dataset compared to its predecessor, focusing on enhancing its roleplay capabilities.

Loading preview...

Overview

Cupid-Qwen3-4B-v0.2 is a 4 billion parameter model developed by Torero96Dev, built upon the Qwen3 architecture. It is a LoRA fine-tune of Goekdeniz-Guelmez/Josiefied-Qwen3-4B-Instruct-2507-gabliterated-v2, with a specific focus on non-reasoning, uncensored roleplay (RP).

Key Capabilities

  • Specialized Roleplay: Designed for engaging in uncensored roleplay scenarios without an emphasis on complex reasoning, aiming for a more fluid and responsive interaction.
  • Optimized for Speed: Developed to offer a faster roleplay experience, contrasting with larger, reasoning-heavy models that can slow down interactions.
  • Enhanced Dataset: This v0.2 iteration includes a cleaned and expanded dataset, improving upon the performance and capabilities of its v0.1 predecessor.

Good For

  • Users seeking a dedicated model for uncensored roleplay.
  • Applications requiring a fast and responsive language model for interactive narrative generation.
  • Developers looking for a Qwen3-based model specifically fine-tuned for creative and non-reasoning conversational tasks.