RMDWLLC/kaiju-coder-7

VISIONConcurrency Cost:2Model Size:27BQuant:FP8Ctx Length:32kTool Calling:SupportedPublished:Jun 3, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

Kaiju Coder 7 by Kiyomi is a 27 billion parameter coding and builder model, fine-tuned by RMDWLLC from Qwen/Qwen3.6-27B with a 32768 token context length. It is specifically optimized for local coding, business-owner build work, and OpenCode-assisted artifact generation. This model excels at generating complete websites, automation scripts, and practical business documents for solo entrepreneurs and local-first AI users.

Loading preview...

Kaiju Coder 7 by Kiyomi: A Specialized Builder Model

Kaiju Coder 7, developed by RMDWLLC and fine-tuned from Qwen/Qwen3.6-27B, is a 27 billion parameter model designed for practical local coding and business-owner artifact generation. It leverages a 32768 token context length, making it suitable for complex tasks. The model's fine-tuning focuses on assisting solo entrepreneurs and local-first AI users in building and automating their businesses.

Key Capabilities

  • Website and Landing Page Generation: Capable of building complete web artifacts.
  • AI-Company Launch Packs: Generates Kiyomi-style launch materials for business owners.
  • Script and Automation Development: Writes scripts, small applications, and automation flows.
  • Business Document Drafting: Produces proposals, launch plans, support notes, and operator handbooks.
  • Data Artifact Creation: Generates intake/CRM schemas, lead lists, ROI dashboards, and reporting artifacts.
  • Architectural Guidance: Helps builders avoid overbuilt solutions and ship useful, practical artifacts.

Intended Use Cases

Kaiju Coder 7 is a strong fit for:

  • Solo-founder product development and small-business website creation.
  • Kiyomi-style local AI product workflows.
  • Practical coding and deployment assistance.

Limitations

It is important to note that Kaiju Coder 7 is not a general frontier model and may be weaker than large cloud models on broad reasoning or uncommon programming domains. It requires a strong harness for advanced tool use and file editing. The model is not intended for high-risk medical, legal, financial, or safety-critical decisions without expert review.