Overview

ModelArts Studio supports the management and deployment of third-party models. Models deployed on ModelArts Studio can be accessed through inference APIs.

**Table 1** APIs
Category	Model	API	Function
Model inference APIs	Third-Party NLP Models	Third-Party NLP Models	DeepSeek API is an API service based on the DeepSeek model. It supports text-based interaction in multiple scenarios and can quickly generate high-quality dialogs, copywriting passages, and stories. It is suitable for scenarios such as text summarization, intelligent Q&A, and content creation.
Model inference APIs	Qwen Third-Party VL Model	Qwen Third-Party VL Model	This Qwen2.5-VL model provides capabilities including image recognition, precise visual positioning, text recognition and understanding, document parsing, and video comprehension.
Data engineering APIs	-	Querying Data Lineages	Allows you to import raw datasets from OBS. You can use the OBS path to query all raw datasets created based on the path and the subsequent lineage information.
Data engineering APIs	-	Permanently Deleting a Dataset	For data uploaded from OBS, you need to delete the associated raw data in OBS when deleting the datasets. Instead of storing raw data in OBS for a long time, customers require that the raw data be archived in their big data center.
Agent APIs	-	Calling an Application	Allows you to call the API of a created application and input a question to obtain the application execution result.
Agent APIs	-	Calling a Workflow	Allows you to call the API of a created workflow and input a question to obtain the workflow execution result.
Token calculator	-	Token Calculator	To help you better manage tokens and optimize token usage, the platform provides the token calculator. The token calculator can help you evaluate the number of tokens in a text before model inference, estimate the cost, and optimize the data preprocessing policy.