dizza01/Qwen2.5-Coder-14B-Instruct

TEXT GENERATIONConcurrency Cost:1Model Size:14.8BQuant:FP8Ctx Length:32kPublished:Apr 30, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

dizza01/Qwen2.5-Coder-14B-Instruct is a 14.7 billion parameter instruction-tuned causal language model from the Qwen2.5-Coder series, developed by Qwen. This model is specifically optimized for code generation, code reasoning, and code fixing, building upon the Qwen2.5 foundation with 5.5 trillion training tokens. It features a full 131,072 token context length and maintains strong performance in mathematics and general competencies, making it suitable for advanced coding applications like Code Agents.

Loading preview...

Qwen2.5-Coder-14B-Instruct Overview

This model is the instruction-tuned 14.7 billion parameter variant of the Qwen2.5-Coder series, developed by Qwen. It represents a significant advancement over CodeQwen1.5, focusing on enhanced coding capabilities while retaining strong general and mathematical competencies. The model's architecture is based on transformers, incorporating RoPE, SwiGLU, RMSNorm, and Attention QKV bias, with 48 layers and 40 attention heads for Q and 8 for KV.

Key Capabilities & Features

  • Code-Specific Optimization: Significantly improved performance in code generation, code reasoning, and code fixing.
  • Extensive Training: Trained on 5.5 trillion tokens, including a substantial amount of source code, text-code grounding, and synthetic data.
  • Long Context Support: Features a full 131,072 token context length, with support for processing even longer texts up to 128K tokens using YaRN for length extrapolation.
  • Foundation for Code Agents: Designed to serve as a comprehensive foundation for real-world applications such as Code Agents, balancing coding prowess with general intelligence.
  • State-of-the-Art Performance: The Qwen2.5-Coder-32B model in this series is noted for matching GPT-4o's coding abilities among open-source code LLMs.

When to Use This Model

This model is particularly well-suited for developers and researchers focused on:

  • Advanced Code Generation: Creating high-quality code across various programming languages.
  • Code Reasoning and Debugging: Tasks requiring logical understanding of code and identifying/fixing errors.
  • Developing Code Agents: Building intelligent systems that can interact with and manipulate code.
  • Applications Requiring Long Code Contexts: Handling large codebases or complex programming tasks that benefit from extended context windows.

Finetunes
1 models