DavidAU/Qwen3.5-9B-Claude-4.6-HighIQ-THINKING-HERETIC-UNCENSORED

VISIONConcurrency Cost:1Model Size:9BQuant:FP8Ctx Length:32kTool Calling:SupportedPublished:Mar 4, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

DavidAU/Qwen3.5-9B-Claude-4.6-HighIQ-THINKING-HERETIC-UNCENSORED is a 9 billion parameter multimodal large language model, fine-tuned from Qwen 3.5 using a Claude 4.6 distill dataset. It features enhanced reasoning capabilities, full uncensored output, and supports vision inputs. This model is optimized for complex thinking generation and real-world adaptability, offering a high degree of instruction following and agentic usage.

Loading preview...

Model Overview

DavidAU/Qwen3.5-9B-Claude-4.6-HighIQ-THINKING-HERETIC-UNCENSORED is a 9 billion parameter multimodal large language model, fine-tuned by DavidAU from the Qwen 3.5 base model. This version significantly enhances the model's reasoning and thinking generation by incorporating a large distill dataset from Claude 4.6. It maintains strong benchmarks while offering uncensored responses and improved instruction following.

Key Capabilities

  • Enhanced Reasoning: VASTLY improved thinking generation, replacing Qwen 3.5's thinking with Claude 4.6's reasoning patterns.
  • Uncensored Output: Functions as a "HERETIC" model, providing responses without content restrictions.
  • Multimodal Support: Vision (image) inputs are fully supported and tested, with video input capabilities inherited from the base Qwen3.5 model.
  • High Instruction Following: Demonstrates strong performance in instruction following tasks, with a refusal rate of only 6/100 compared to the original model's 100/100.
  • Agentic Usage: Excels in tool calling capabilities, with recommended integration via Qwen-Agent and Qwen Code for building agent applications.
  • Extended Context: Natively supports a context length of 262,144 tokens, extensible up to 1,010,000 tokens using YaRN scaling techniques.

Good For

  • Applications requiring advanced reasoning and complex problem-solving.
  • Use cases where uncensored and direct responses are critical.
  • Multimodal tasks involving image and video understanding.
  • Developing AI agents that require robust tool-calling and instruction following.

Finetunes
1 models