Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled · Hugging Face

Source: https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%8C%9F-qwen35-27b-claude-46-opus-reasoning-distilled🌟 Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled

🔥**Update (April 5):**I’ve released the complete training notebook, codebase, and a comprehensive PDF guide to help beginners and enthusiasts understand and reproduce this model’s fine-tuning process.

❤️ Special thanks to theUnslothopen-source library and@KyleHessling1for their support.

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%93%9A-resources–guides📚 Resources & Guides

👉**GitHub Repository: Jackrong-llm-finetuning-guide**Visit the repo to dive into the codebase and reproduce the results locally or on Colab.

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%93%A5-core-technical-document📥 Core Technical Document

🔗Qwopus3.5-27b Complete Fine-Tuning Guide (PDF)

• **The Full Pipeline:**A step-by-step walkthrough—from downloading the base model and unifying heterogeneous data, to configuring trainer hyperparameters and publishing to Hugging Face.
• **Beginner Friendly:**Includes an introductory guide to getting started with Google Colab and Unsloth.
• Feedback welcome! If you spot any areas for improvement, please let me know and I will update it promptly.

A Note:My goal isn’t just to detail a workflow, but to demystify LLM training. Beyond the social media hype, fine-tuning isn’t an unattainable ritual—often, all you need is a Google account, a standard laptop, and relentless curiosity. No one starts as an expert, but every expert was once brave enough to begin. All training and testing for this project were self-funded. If you find this model or guide helpful, aStar ⭐️ on GitHubwould be the greatest encouragement. Thank you! 🙏

The Claude series model optimizations are named under theQwopus3.5 series, with the latest version being**🌟Qwopus3.5-v3**.

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%8C%9F-qwen35-27b-claude-46-opus-reasoning-distilled-1🌟 Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled

Build Environment Upgrades: - Fine-tuning Framework:Unsloth 2026.3.3 - Core Dependencies:Transformers 5.2.0 - This model fixes the crash in the official model caused by the Jinja template not supporting the**“developer”role. (commonly sent by modern coding agents like Claude Code and OpenCode) - It doesnot disable thinking mode by default**, and allowing the agent to run continuously forover 9 minutes without interruption. - Compared to the original model,autonomy and stability are significantly improved.

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%92%A1-model-introduction💡 Model Introduction

Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilledis a highly capable reasoning model fine-tuned on top of the powerful Qwen3.5 architecture. The model’s core directive is to leverage state-of-the-art Chain-of-Thought (CoT) distillation primarily sourced from Claude-4.6 Opus interactions.

Through Supervised Fine-Tuning (SFT) focusing specifically on structured reasoning logic, this model excels in breaking down complex user problems, planning step-by-step methodologies within strictly formatted<think\>tags, and ultimately delivering precise, nuanced solutions.

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%A7%A0-example-of-learned-reasoning-scaffold%EF%BC%88example%EF%BC%89🧠 Example of Learned Reasoning Scaffold（Example）

The model includes targeted optimizations addressing Qwen3.5’s tendency toward excessive transitional or repetitive reasoning on simple queries. Through deep distillation and structural imitation of Claude-4.6-Opus reasoning chains, the model adopts a more efficient structured thinking pattern: “Let me analyze this request carefully: 1..2..3...”. This streamlined reasoning paradigm significantly reduces redundant cognitive loops while preserving deep analytical capacity, resulting in substantially improved inference efficiency.

Let me analyze this request carefully:

1. Identify the core objective of the problem.
2. Break the task into clearly defined subcomponents.
3. Evaluate constraints and edge cases.
4. Formulate a step-by-step solution plan.
5. Execute the reasoning sequentially and verify consistency.
            .
            .
            .

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%97%BA%EF%B8%8F-training-pipeline-overview🗺️ Training Pipeline Overview

Base Model (Qwen3.5-27B)
 │
 ▼
Supervised Fine-Tuning (SFT) + LoRA
 │
 ▼
Final Model (Claude-4.6-Opus-Reasoning-Distilled,text-only)

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%93%8B-stage-details📋 Stage Details

🔧Tool Calling Benchmark（benchmark tests by user @Chris Klaus）

From the test results, it is clear that different Qwen3.5 quantized models show significant differences in tool-calling capability. Among them, only the 27B model distilled with Claude Opus reasoning demonstrates stable performance.

🔥Community-tested advantages(benchmark tests by user @sudoing on a single RTX 3090):

Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled shows significant advantages in coding-agent environments such as Claude Code and OpenCode:

• Native support for the “developer” role, requiring no Jinja template patches or ChatML workarounds. - Thinking mode fully preserved(logs confirmthinking=1), not silently disabled, maintaining the complete chain-of-thought reasoning process. - Greatly improved autonomy and stability— capable of running continuously forover 9 minutes autonomously(with zero human intervention). It actively waits for tool responses, reads outputs, self-corrects errors, and can even automatically generate a README, whereas the base model often stalls or freezes mid-execution.

Hardware usage remains unchanged: - About16.5 GB VRAMwithQ4_K_Mquantization - 29–35 tok/sgeneration speed - Full 262K contextwith no compromises

• These improvements come from successfully distilling thestructured reasoning style of Claude 4.6 Opus, allowing Qwopus to be trulyplug-and-play in modern local coding agentsand deliver an experience close to Opus in smoothness and usability.

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%94%B9-supervised-fine-tuning-sft🔹 Supervised Fine-Tuning (SFT)

• **Objective:**To inject high-density reasoning logic and establish a strict format for problem-solving involving an internal thinking state prior to outputting the final response.
• Methodology:We utilizedUnslothfor highly efficient memory and compute optimization. A critical component of this stage is thetrain\_on\_responses\_onlystrategy, masking instructions so the loss is purely calculated over the generation of the<think\>sequences and the subsequent solutions.
• **Format Enforcement:**All training samples were systematically normalized so the model strictly abides by the structure<think\> \{internal reasoning\} </think\>\\n \{final answer\}.

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%93%9A-all-datasets-used📚 All Datasets Used

The dataset consists of high-quality, filtered reasoning distillation data:

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%8C%9F-core-skills–capabilities🌟 Core Skills & Capabilities

1. **Modular & Structured Thinking:**Inheriting traits from Opus-level reasoning, the model demonstrates confident parsing of the prompt, establishing an outlined plan in its<think\>block sequentially rather than exploratory “trial-and-error” self-doubt.

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%E2%9A%A0%EF%B8%8F-limitations–intended-use⚠️ Limitations & Intended Use

• **Hallucination Risk:**While reasoning is strong, the model remains an autoregressive LLM; external facts provided during the thinking sequence may occasionally contain hallucinations if verifying real-world events.
• **Intended Scenario:**Best suited for offline analytical tasks, coding, math, and heavy logic-dependent prompting where the user needs to transparently follow the AI’s internal logic.
• **Preview Version Notice:**Because this model is relatively new and intentionally lightweight, the surrounding ecosystem — including inference templates, fine-tuning pipelines, routing configurations, and tooling integrations — may not yet be fully mature or standardized. As a result, users may encounter occasional bugs, compatibility inconsistencies, or integration edge cases. The current release should be considered a preview build while the broader architectural stack and supporting utilities continue to stabilize and improve.

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%99%8F-acknowledgements🙏 Acknowledgements

Significant thanks to theUnsloth AIteam for making rapid fine-tuning of MoE and large LLM models accessible. Additionally, we acknowledge Qwen internally, and the open-source community developers producing exceptional distilled datasets (nohurryandTeichAI).

https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled#%F0%9F%93%96-citation📖 Citation

If you use this model in your research or projects, please cite:

@misc{jackrong_qwen35_opus_distilled,
  title        = {Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled},
  author       = {Jackrong},
  year         = {2026},
  publisher    = {Hugging Face},
  howpublished = {\url{https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled}}
}