You can integrate plug-ins with large language model (LLM) applications to enhance their functionality against complex tasks, such as obtaining the latest information, avoiding hallucinations, and perform precise calculations.
Overview
A plug-in is a collection of one or more tools (APIs), with each tool providing a specific feature. Model Studio supports official plug-ins and custom plug-ins.
Official plug-ins
The Plug-in Gallery offers pre-installed official plug-ins from Model Studio. You can use them directly without configuring input and output parameters, see Official plug-ins.
Name | Tool ID | Description | Billing |
code_interpreter | Allows the LLM to execute Python code snippets, such as mathematical calculations, data analysis and visualization, and data processing. | Free | |
calculator | Allows the LLM to perform complex mathematical calculations, such as “12313×13232”. | Free | |
text_to_image | Allows the LLM to generate images based on text, such as “please draw a smiling puppy”. | Free for a limited period. You must apply for the plug-in first. |
Custom plug-ins
You can create custom plug-ins and integrate them into your LLM applications for flexible and personalized feature expansion.
Performance showcase
Sample input | Without a plug-in | With a plug-in |
What is 12313×13232? | The application cannot accurately solve the problem and may return incorrect answers. In this case, the correct answer is 162,925,616. | The application now has robust calculation capability and can generate accurate calculation results. |
|
|
Supported models
Model | Model identifier |
Qwen-Turbo | qwen-turbo |
Qwen-Plus | qwen-plus |
Qwen-Max | qwen-max |
Qwen VL-Max | qwen-vl-max |
Qwen VL-Plus | qwen-vl-plus |
Qwen2-57B-A14B-Instruct | qwen2-57b-a14b-instruct |
Qwen2-72B-Instruct | qwen2-72b-instruct |
The open source Qwen2 models are not currently compatible with the Image Generation plug-in. Refer to the actual execution results in the console for the latest compatibility status.
For more information about the models, see the List of models.
How plug-in works
Calling a plug-in essentially means calling the tools within it. Model Studio supports the calling of plug-ins through agent applications and workflow applications.
When a plug-in is called in an agent applications, the LLM determines whether to call the tools based on the user input, tool name, and tool description.
If a tool needs to be called, the LLM selects the appropriate tool. After the tool is called within the application, the tool's output is merged with the user input and re-entered into the LLM, which generates and outputs the final result.
If no tool needs to be called, the LLM generates and outputs the result directly.
In workflow applications, the plug-in acts as a node and executes tasks according to the user's orchestration, rather than being actively planned and called by the LLM.