Qwen2.5-Coder-32B-Instruct
Code generation model Generally available Open WeightsAlibaba's flagship 32B open-weights code model featuring Apache 2.0 licensing, 128k context, and multi-language programming mastery.
Overview
Qwen2.5-Coder-32B-Instruct is Alibaba's flagship open-weights code language model released in November 2024. Built with 32.5 billion parameters, it achieves coding benchmark performance comparable to leading proprietary frontier models across Python, C++, Java, JavaScript, TypeScript, Go, and Rust.
The model features a 131,072-token (128k) context window with up to 8,192 output tokens. It supports repository-level code comprehension, automated bug fixing, unit test generation, and structured tool calling.
Published under the Apache 2.0 open-source license, Qwen2.5-Coder-32B-Instruct is downloadable for local deployment (via Ollama, vLLM, and LM Studio) and is hosted via Alibaba Cloud Model Studio API ($0.20 / 1M input tokens, $0.60 / 1M output tokens).
Technical Specifications
Context Window
131,072 tokens
Max Input Tokens
32,768
Max Output Tokens
8,192
Parameters
32.5B
Architecture
Transformer
Released
2024-09-19
Modalities
Text → Text (Both)
Capabilities
Text Generation
Code Generation
Function Calling
Use Cases
Coding Assistance
Code Review
Languages
English (en)
Chinese (zh)
Access Routes
| Route | Type | Protocol | Region / Scope | Provider | Status |
|---|---|---|---|---|---|
| Alibaba Cloud | Hosted API (Direct provider) | OpenAI-compatible REST API | Alibaba Cloud | Available |
Identifiers
- qwen2.5-coder-32b-instruct Canonical identifier Default
- Qwen/Qwen2.5-Coder-32B-Instruct Repository identifier Default