icecee/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-Fix

VISIONPricing:Input $0.431 / Cached $0.0862 / Output $1.12Concurrent Unit Cost:1Model Size:9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Mar 11, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

icecee/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-Fix is a 9-billion parameter language model built on the Qwen3.5-9B architecture, fine-tuned for advanced reasoning. It leverages Chain-of-Thought (CoT) distillation from Claude-4.6 Opus interactions to excel at breaking down complex problems and generating structured, step-by-step solutions. This model is optimized for tasks requiring transparent internal logic, such as analytical problem-solving, coding, and mathematics, by enforcing a strict tag format. It features an extended context window of 16,384 tokens, supporting complex multi-step reasoning traces.

Loading preview...

Model Overview

icecee/Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-Fix is a 9-billion parameter model based on the Qwen3.5-9B architecture, specifically fine-tuned for enhanced reasoning capabilities. Its core strength lies in leveraging Chain-of-Thought (CoT) distillation, primarily sourced from Claude-4.6 Opus interactions, to process complex problems.

Key Capabilities & Features

  • Structured Reasoning: The model excels at breaking down complex user problems and planning step-by-step methodologies within strictly formatted <think> tags, delivering precise and nuanced solutions.
  • Efficient Thinking: It adopts a streamlined reasoning paradigm, reducing redundant cognitive loops while preserving deep analytical capacity, leading to improved inference efficiency.
  • Distilled Knowledge: Training involved Supervised Fine-Tuning (SFT) using high-quality reasoning distillation data from datasets like nohurry/Opus-4.6-Reasoning-3000x-filtered and TeichAI/claude-4.5-opus-high-reasoning-250x.
  • Extended Context: Supports an extended context window of 16,384 tokens, allowing for complex multi-step reasoning traces.
  • Enhanced Reasoning Data: Further improved with additional reasoning data distilled from Qwen3.5-27B, including Jackrong/Qwen3.5-reasoning-700x, to introduce higher-quality reasoning trajectories across science, instruction-following, and mathematics.

Intended Use Cases

This model is best suited for scenarios requiring transparent internal logic and structured problem-solving:

  • Offline analytical tasks
  • Coding and mathematical problem-solving
  • Heavy logic-dependent prompting where the user needs to follow the AI's internal thought process.

Finetunes
1 models