紫东太初

Overview

Zidong Taichu is a multimodal large model jointly launched by the Institute of Automation, Chinese Academy of Sciences and the Wuhan Institute of Artificial Intelligence, and has now been upgraded to version 2.0. The product provides conversational AI services to the public and supports tasks such as multi-turn Q&A, text creation, image generation, video understanding, audio analysis, 3D scene understanding, and signal recognition, making it suitable for comprehensive Q&A, content generation, and multimodal interaction scenarios.

Compared with traditional text assistants, Zidong Taichu emphasizes “full-modality” capabilities. Users can not only enter text, but also upload images, videos, audio, music, point clouds, and signal files for targeted Q&A and analysis.

Official website: https://taichu-web.ia.ac.cn/#/welcome

Main Features

• Text and Dialogue

  • Supports Chinese Q&A, multi-turn dialogue, text continuation, article writing, and title generation
  • Provides grammatical analysis, machine translation, classical poetry creation, mathematical calculation, and logical reasoning
  • Supports code understanding and simple code writing

• Image Capabilities

  • Can perform image description, object detection, image retrieval, and image generation
  • Supports text recognition based on image content, covering OCR needs across multiple scenarios and languages

• Video Capabilities

  • Supports video description, video retrieval, and video Q&A
  • Can combine context for continuous follow-up questions, making it suitable for video content understanding scenarios

• Audio and Music Capabilities

  • Supports speech recognition, speech synthesis, audio forgery detection, and audio event classification
  • Can generate music based on text prompts, and understand uploaded music content to complete related Q&A

• 3D and Signal Capabilities

  • Supports 3D scene description and object perception based on point cloud data
  • Supports signal recognition such as radar and related knowledge interaction

Product Pricing

At present, the publicly available information on the official website does not clearly display a standardized pricing plan. According to the existing information, users usually need to register an account first and wait for review. After approval, they can enter the dialogue interface to try it out. For the latest usage rules and scope of availability, please refer to the official website.

FAQ

How do I use Zidong Taichu?

• Visit the official website and click the dialogue experience
• Register or log in to your account
• Enter the dialogue interface after approval
• Enter your question, or use the prompts and examples provided by the system to start an interaction

What file types are supported for upload?

Zidong Taichu supports uploading files such as images, videos, point clouds, audio, music, and signals, and can conduct Q&A and analysis around the uploaded content.

Has it completed generative AI filing?

According to public information, Zidong Taichu was among the first batch to pass the filing under the Interim Measures for the Administration of Generative Artificial Intelligence Services in August 2023, and can officially provide services to the public.