Qwen/Qwen2.5-Coder-0.5B-Instruct

Qwen2.5-Coder-0.5B-Instruct is a 0.49 billion parameter instruction-tuned causal language model developed by Qwen, part of the Qwen2.5-Coder series. This model is specifically optimized for code generation, code reasoning, and code fixing, building upon the Qwen2.5 architecture. It features a full context length of 32,768 tokens and is designed for real-world coding applications, including Code Agents, while maintaining general and mathematical competencies.