VINAY-UMRETHE/Qwen3-0.6B-heretic-Test5

TEXT GENERATIONConcurrency Cost:1Model Size:0.8BQuant:BF16Ctx Length:32kTool Calling:SupportedPublished:Jun 4, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

VINAY-UMRETHE/Qwen3-0.6B-heretic-Test5 is a 0.8 billion parameter causal language model, derived from Qwen/Qwen3-0.6B, and decensored using Heretic v1.3.0. This model features a 32,768 token context length and is specifically optimized for reduced refusals compared to its original counterpart. It excels in scenarios requiring less restrictive content generation while retaining the Qwen3 architecture's capabilities in reasoning, instruction-following, and multilingual support.

Loading preview...

Model Overview

VINAY-UMRETHE/Qwen3-0.6B-heretic-Test5 is a 0.8 billion parameter causal language model, built upon the Qwen3 architecture and decensored using Heretic v1.3.0. It maintains a substantial context length of 32,768 tokens, making it suitable for processing lengthy inputs and generating comprehensive responses. A key differentiator of this model is its significantly reduced refusal rate, achieving 12 refusals out of 100 compared to the original model's 55/100, making it more permissive in content generation.

Key Capabilities

  • Decensored Content Generation: Offers a less restrictive output compared to the base Qwen3-0.6B model.
  • Flexible Thinking Modes: Inherits Qwen3's unique ability to seamlessly switch between 'thinking' (for complex reasoning, math, coding) and 'non-thinking' (for efficient general dialogue) modes, enhancing performance across diverse tasks.
  • Enhanced Reasoning: Demonstrates strong capabilities in mathematics, code generation, and commonsense logical reasoning.
  • Multilingual Support: Supports over 100 languages and dialects, with robust multilingual instruction following and translation abilities.
  • Agentic Functionality: Excels in tool-calling and integration with external tools, performing well in complex agent-based tasks.

Good For

  • Applications requiring less content moderation: Ideal for use cases where the base model's refusal rates are too high.
  • Complex problem-solving: Benefits from the 'thinking mode' for tasks involving logical reasoning, mathematical computations, and code generation.
  • Creative writing and role-playing: Offers superior human preference alignment for engaging conversational experiences.
  • Multilingual interactions: Suitable for global applications needing strong instruction following and translation across many languages.