Model Selection

VJSP currently offers the following four dedicated models. The key to generating effective code is matching the right mode and model to your specific task. Below is an overview of each model’s purpose and ideal use cases:

• Qwen3-Coder:

Supports code generation across various programming languages, suitable for full-spectrum development scenarios. It delivers efficient and precise coding assistance. Use this model when generating general-purpose code.

• Qwen3-Thinking:

Specializes in complex logical reasoning and multi-step problem solving, ideal for highly challenging tasks that require deep decomposition and step-by-step deduction.

• Qwen3-VJSP:

Focused on understanding and handling development tasks related to the VJSP full-stack framework—including framework component usage, front-end/back-end interaction logic, business feature implementation, configuration file authoring, and full-stack deployment strategies. Use this model when generating VJSP framework-specific code.

• Qwen3-VL:

A specialized AI model designed for processing and understanding image data. Use this model when generating VJSP workflows involving visual inputs.