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

Overview

DavidAU/Qwen3.5-9B-Claude-4.6-HighIQ-INSTRUCT is a 9 billion parameter instruction-tuned model based on the Qwen 3.5 architecture. It was fine-tuned using a Claude 4.6 large distill dataset, with a focus on maintaining the strong benchmarks of the base model while enhancing its instruction-following capabilities. The model supports a native context length of 262,144 tokens, extensible up to 1,010,000 tokens using YaRN scaling.

Key Capabilities

• Enhanced Reasoning & Instruction Following: Benchmarks show significant improvements over the base Qwen3.5-9B model in tasks like ARC, BoolQ, HSwag, OBQA, PIQA, and Wino. It also excels in instruction following benchmarks like IFEval and MultiChallenge.
• Multimodal Understanding: The model is vision-capable, with tested image input functionality. The base Qwen3.5 model also supports video input, though this specific fine-tune's video capabilities were not explicitly tested by the developer.
• Agentic Usage: Optimized for tool calling and agent applications, with recommendations for use with Qwen-Agent and Qwen Code.
• Long Context Handling: Natively supports a 262,144 token context, with methods to extend this to over 1 million tokens for ultra-long text processing.

When to Use This Model

• Complex Reasoning Tasks: Ideal for applications requiring advanced logical deduction, problem-solving, and high-IQ responses.
• Instruction-Following Applications: Suited for scenarios where precise adherence to instructions is critical.
• Multimodal Interactions: Useful for tasks involving image understanding and potentially video analysis (though video capabilities require further testing for this specific fine-tune).
• Agent Development: A strong candidate for building AI agents that require robust tool-calling and environmental interaction capabilities.