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

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 is a 9 billion parameter Qwen 3.5 dense model fine-tuned by DavidAU using a Claude 4.6 large distill dataset. This model significantly enhances the thinking generation capabilities of the base Qwen 3.5, replacing its native thinking with that of Claude 4.6. It maintains strong original benchmarks and supports vision inputs, making it suitable for complex reasoning tasks and multimodal applications.

Loading preview...

Model Overview

DavidAU/Qwen3.5-9B-Claude-4.6-HighIQ-THINKING is a 9 billion parameter model based on the Qwen 3.5 architecture, fine-tuned by DavidAU. The core differentiator of this model is its enhanced "thinking generation" capabilities, achieved by distilling knowledge from a Claude 4.6 dataset. This process aims to imbue the model with Claude 4.6's reasoning patterns while preserving the strong performance of the base Qwen 3.5 model.

Key Capabilities

  • Enhanced Reasoning: Significantly improved thinking generation, as evidenced by benchmark improvements over the base Qwen 3.5 model in tasks like ARC, BoolQ, HSwag, OBQA, PIQA, and Wino.
  • Multimodal Support: Inherits and maintains vision capabilities from the Qwen 3.5 base model, with tested working image processing.
  • High Context Length: Supports a native context length of 262,144 tokens, extensible up to 1,010,000 tokens using YaRN scaling techniques.
  • Efficient Architecture: Leverages Qwen 3.5's efficient hybrid architecture, including Gated Delta Networks and sparse Mixture-of-Experts, for high-throughput inference.

Good For

  • Complex Reasoning Tasks: Ideal for applications requiring advanced logical thought processes and problem-solving, benefiting from the Claude 4.6-derived thinking.
  • Multimodal Applications: Suitable for tasks involving both text and image inputs, such as visual question answering or image analysis.
  • Long Context Processing: Excellent for handling extensive documents or conversations due to its large native and extensible context window.
  • Agentic Workflows: Designed to excel in tool calling and agent applications, with specific recommendations for Qwen-Agent and Qwen Code integration.

Finetunes
2 models