Overview

MagicalAlchemist/Qwen3-1.7B-Magic_decensored is a 1.7 billion parameter causal language model, derived from Qwen/Qwen3-1.7B and decensored using Heretic v1.1.0. This modification significantly reduces the model's refusal rate from 65/100 to 45/100, making it more permissive for diverse use cases. It maintains the original Qwen3 architecture, offering a substantial 40960 token context length and advanced capabilities in reasoning, instruction-following, and agentic tasks.

Key Capabilities

• Decensored Behavior: Achieves a lower refusal rate (45/100) compared to the original model (65/100).
• Dual Thinking Modes: Supports seamless switching between a 'thinking mode' for complex logical reasoning, math, and coding, and a 'non-thinking mode' for efficient general-purpose dialogue.
• Enhanced Reasoning: Demonstrates improved performance in mathematics, code generation, and commonsense logical reasoning.
• Superior Human Preference Alignment: Excels in creative writing, role-playing, multi-turn dialogues, and instruction following.
• Agentic Expertise: Integrates precisely with external tools, achieving leading performance in complex agent-based tasks among open-source models.
• Multilingual Support: Capable of handling over 100 languages and dialects with strong multilingual instruction following and translation abilities.

Good For

• Applications requiring a more permissive language model with reduced content restrictions.
• Tasks demanding advanced logical reasoning, mathematical problem-solving, and code generation.
• Creative writing, role-playing, and engaging multi-turn conversational agents.
• Complex agentic workflows and tool integration.
• Multilingual applications, instruction following, and translation across numerous languages.