One of the most important aspects of AI projects is being able to quickly experiment with different models and easily integrate them into existing systems. A flexible, accessible, and fast-action structure provides a great advantage in this process.
In this article, we will briefly look at Huawei Cloud's Model-as-a-Service (MaaS) offering and explore how we can access AI models and use them quickly through a sample implementation.
What is MaaS?
MaaS (Model-as-a-Service) allows us to access different AI models from a single platform on Huawei Cloud. It enables us to use ready-made models directly through APIs without setting up infrastructure.
On the MaaS side, DeepSeek, GLM, and Qwen model families are currently available, and we can use these models directly for different scenarios.
Setup
We can see available models by going to the MaaS service details on the console.
By clicking Call on the model, we can see how to use it as shown in the image below. We can send requests in Standard and OpenAI-compatible formats. To make requests, we need to create an API Key.
We can activate the models we want to use from Real-Time Inference.
Test
