DavidAU/Qwen3.5-9B-Claude-4.6-OS-Auto-Variable-THINKING

VISIONConcurrent Unit Cost:1Model Size:9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Mar 6, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

DavidAU/Qwen3.5-9B-Claude-4.6-OS-Auto-Variable-THINKING is a 9 billion parameter Qwen 3.5 dense model fine-tuned by DavidAU using the Claude-4.6-OS dataset. This model features altered reasoning and thinking block sizes, aiming to enhance performance without negatively impacting its strong benchmarks. It is a vision-capable model with a native context length of 262,144 tokens, extensible up to 1,010,000 tokens, and is optimized for complex reasoning tasks and multimodal understanding.

Loading preview...

Model Overview

DavidAU/Qwen3.5-9B-Claude-4.6-OS-Auto-Variable-THINKING is a 9 billion parameter model based on the Qwen 3.5 architecture, fine-tuned by DavidAU. This model incorporates a Claude-4.6-OS dataset and features modifications to its reasoning and thinking block sizes. It is designed to maintain the strong performance of the base Qwen 3.5 model while potentially improving its reasoning capabilities.

Key Capabilities

  • Enhanced Reasoning: The model has altered reasoning/thinking blocks, aiming for improved performance in complex tasks.
  • Multimodal Support: It is a vision-capable model, supporting image and video inputs, with video portions confirmed to be working with new training.
  • Extended Context Length: Natively supports a context length of 262,144 tokens, extensible up to 1,010,000 tokens using YaRN scaling techniques.
  • Tool Calling: Excels in tool calling capabilities, with recommended integration via Qwen-Agent and Qwen Code.
  • Benchmark Performance: Shows competitive performance across various benchmarks, including MMLU-Pro (82.5), IFEval (91.5), MMMU (78.4), and CountBench (97.2).

When to Use This Model

  • Complex Reasoning Tasks: Ideal for applications requiring advanced reasoning, especially where Claude-like thinking processes are beneficial.
  • Multimodal Applications: Suitable for tasks involving image and video understanding, such as visual question answering or video summarization.
  • Long Context Processing: Excellent for use cases requiring the processing of ultra-long texts, leveraging its extended context window.
  • Agentic Workflows: Recommended for building AI agents, particularly with Qwen-Agent or Qwen Code, due to its strong tool-calling abilities.

Finetunes
1 models