kkomyoeminaung/Instruct-and-coder-merged

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:7.6BQuant:FP8Ctx Length:32kPublished:May 27, 2026Architecture:Transformer Warm

The kkomyoeminaung/Instruct-and-coder-merged is a 7.6 billion parameter language model, merged from Qwen/Qwen2.5-7B-Instruct and Qwen/Qwen2.5-Coder-7B-Instruct using a linear merge method. With a 32768 token context length, this model is designed to combine general instruction following capabilities with enhanced code generation and understanding. It is particularly suited for tasks requiring both natural language interaction and programming assistance.

Loading preview...

Model Overview

The kkomyoeminaung/Instruct-and-coder-merged is a 7.6 billion parameter language model created by kkomyoeminaung. It was developed using the mergekit tool, specifically employing a linear merge method to combine the strengths of two base models from the Qwen family.

Merge Details

This model is a blend of:

  • Qwen/Qwen2.5-7B-Instruct: This model forms the primary base, contributing 80% of the merged model's characteristics.
  • Qwen/Qwen2.5-Coder-7B-Instruct: This model contributes 20%, enhancing the merged model's coding capabilities.

The linear merge strategy was applied to integrate these components, aiming to create a versatile model with a 32768 token context length that excels in both general instruction following and specialized code-related tasks.

Key Capabilities

  • Instruction Following: Inherits robust general instruction-following abilities from its Qwen2.5-7B-Instruct base.
  • Code Generation & Understanding: Benefits from the Qwen2.5-Coder-7B-Instruct component, providing enhanced performance in coding tasks.
  • Versatile Application: Suitable for use cases requiring a balance of natural language processing and programming assistance.

When to Use This Model

This model is ideal for developers and researchers who need a single model capable of handling a wide range of tasks, from conversational AI and content generation to code completion, debugging, and explanation. Its merged architecture aims to provide a balanced performance across these domains, making it a strong candidate for applications that require both strong instructional understanding and coding proficiency.