xtools-at/Qwen3.5-4B-Claude-3xMerge-Jackrong

VISIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:4.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Oct 12, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

xtools-at/Qwen3.5-4B-Claude-3xMerge-Jackrong is a 4.5 billion parameter language model developed by Jackrong, created by merging three Qwen3.5-4B fine-tunes. This experimental merge combines models optimized for Claude Opus-style reasoning distillation, coding, and structured multi-step reasoning. It is primarily focused on enhancing reasoning, programming, debugging, and agent-oriented workflows, offering a compact solution for complex problem-solving tasks.

Loading preview...

Overview

This model, xtools-at/Qwen3.5-4B-Claude-3xMerge-Jackrong, is an experimental 4.5 billion parameter checkpoint merge based on the Qwen/Qwen3.5-4B architecture. Developed by Jackrong, it combines three distinct fine-tuned models to create a single, compact solution with enhanced capabilities. The merge specifically targets improved reasoning, coding, and general problem-solving, with additional focus on structured reasoning and agentic tool-use workflows.

Key Capabilities

  • Enhanced Reasoning: Distilled from Claude Opus 4.6-style traces, focusing on structured problem decomposition, logical reasoning, and efficient analytical problem-solving.
  • Advanced Coding: Optimized for software development workflows, including code generation, debugging, and tool-assisted problem-solving through trace inversion and agent trajectories.
  • Structured Problem-Solving: Emphasizes explicit multi-step reasoning, complex problem decomposition, and consistent reasoning across tasks.
  • Agentic Workflows: Designed to support experimentation with structured tool calls and iterative execution in compatible environments.

Good For

  • General Reasoning Tasks: Ideal for multi-step questions, logical deduction, and analytical challenges.
  • Programming & Debugging: Suitable for generating code, explaining existing code, and identifying bugs.
  • Technical Problem-Solving: Effective in breaking down complex technical tasks and evaluating solutions.
  • Agent Development: Useful for exploring and implementing agent-oriented workflows with structured tool use.
  • Local Inference: Designed for evaluation and deployment in environments supporting the Qwen3.5 architecture, offering a compact model for advanced tasks.

Finetunes
3 models