Learn about the currently available large language models (LLMs) and how to change the default model.
Make sure your Lingma version is 2.1.0 or later in the IDE. See Install Lingma for details.
Models
The following models are currently available:
qwen3
qwen3-thinking
qwen2.5-max
In the Enterprise Dedicated edition, you can choose the self-hosted models that are configured in the console.
How to change models
By default, Lingma uses the Qwen LLMs. Here's how to change it to another LLM:
Go to the IDE and open Lingma.
In the left of the input box, click the model name and select a model from the drop-down menu.
Aspects to keep in mind
The DeepSeek models are comprehensive model services hosted on the Alibaba Cloud Bailian platform.
To ensure optimal response quality, Qwen LLMs will be used if the model you select does not support a particular scenario. For example, if deepseek-v3 is incompatible with multimodal tasks, Lingma will use Qwen for chats with images.