Jackrong/Qwen3.5-2B-Claude-4.6-Opus-Reasoning-Distilled

VISIONConcurrency Cost:1Model Size:2.3BQuant:BF16Ctx Length:32kTool Calling:SupportedPublished:Mar 3, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

Jackrong/Qwen3.5-2B-Claude-4.6-Opus-Reasoning-Distilled is a 2.3 billion parameter language model built on the Qwen3.5-2B architecture, fine-tuned for advanced reasoning. It leverages Chain-of-Thought (CoT) distillation from Claude-4.6 Opus interactions, focusing on structured step-by-step problem-solving within a 32768-token context window. This model excels at breaking down complex problems, planning methodologies, and delivering precise solutions, making it ideal for analytical tasks, coding, and mathematics.

Loading preview...

Qwen3.5-2B-Claude-4.6-Opus-Reasoning-Distilled Overview

This model is a 2.3 billion parameter language model, built upon the Qwen3.5-2B base, and specifically fine-tuned for enhanced reasoning capabilities. It incorporates advanced Chain-of-Thought (CoT) distillation techniques, primarily sourced from Claude-4.6 Opus interactions, to foster structured, step-by-step problem-solving.

Key Enhancements & Capabilities

  • Reasoning Distillation: Further enhanced with high-quality reasoning trajectories distilled from Qwen3.5-27B, improving performance in science, instruction-following, and mathematics.
  • Structured Thinking: Employs a streamlined reasoning paradigm, adopting an efficient structured thinking pattern (e.g., "Let me analyze this request carefully: 1..2..3...") to reduce redundant cognitive loops.
  • SFT with Unsloth: Utilizes Supervised Fine-Tuning (SFT) with Unsloth for memory and compute optimization, focusing on training the model to generate internal <think> sequences before producing final answers.
  • Extended Context: Supports an extended context window of 16,384 tokens, allowing for complex multi-step reasoning within memory limits.
  • Dataset Integration: Trained on curated datasets like nohurry/Opus-4.6-Reasoning-3000x-filtered and Jackrong/Qwen3.5-reasoning-700x to strengthen structured problem-solving and reasoning diversity.

Intended Use Cases

This model is best suited for offline analytical tasks, coding, mathematical problem-solving, and other logic-dependent prompting scenarios where transparent, step-by-step internal logic is beneficial. It is a test version for academic research and technical exploration.